Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Post-glacial rebound wikipedia , lookup
Cascade Volcanoes wikipedia , lookup
Great Lakes tectonic zone wikipedia , lookup
Algoman orogeny wikipedia , lookup
Geology of the Pyrenees wikipedia , lookup
Plate tectonics wikipedia , lookup
Large igneous province wikipedia , lookup
Mantle plume wikipedia , lookup
Cimmeria (continent) wikipedia , lookup
JOURNAL OF GEOPHYSICAL RESEARCH: SOLID EARTH, VOL. 118, 5530–5549, doi:10.1002/jgrb.50390, 2013 Downdip landward limit of Cascadia great earthquake rupture R. D. Hyndman 1,2 Received 31 January 2013; revised 19 September 2013; accepted 29 September 2013; published 24 October 2013. [1] This paper examines the constraints to the downdip landward limit of rupture for the Cascadia great earthquakes off western North America. This limit is a primary control for ground motion hazard at near-coastal cities. The studies also provide information on the physical controls of subduction thrust rupture globally. The constraints are (1) “locked/ transition” zones from geodetic deformation (GPS, repeated leveling, tide gauges); (2) rupture zone from paleoseismic coastal marsh subsidence, “paleogeodesy”; (3) temperature on the thrust for the seismic-aseismic transition; (4) change in thrust seismic reflection character downdip from thin seismic to thick ductile; (5) fore-arc mantle corner aseismic serpentinite and talc overlying the thrust; (6) updip limit of episodic tremor and slip (ETS) slow slip; (7) rupture area associations with shelf-slope basins; (8) depth limit for small events on the thrust; and (9) landward limit of earthquakes on the Nootka transform fault zone. The most reliable constraints for the limit of large rupture displacement, >10 m, are generally just offshore in agreement with thermal control for this hot subduction zone, but well-offshore central Oregon and near the coast of northern Washington. The limit for 1–2 m rupture that can still provide strong shaking is less well estimated 25–50 km farther landward. The fore-arc mantle corner and the updip extent of ETS slow slip are significantly landward from the other constraints. Surprisingly, there is a downdip gap between the best other estimates for the great earthquake rupture zone and the ETS slow slip. In this gap, plate convergence may occur as continuous slow creep. Citation: Hyndman, R. D. (2013), Downdip landward limit of Cascadia great earthquake rupture, J. Geophys. Res. Solid Earth, 118, 5530–5549, doi:10.1002/jgrb.50390. 1. Introduction [2] The downdip landward limit of rupture for the Cascadia subduction zone off western North America is a primary control for the ground shaking hazard from great earthquakes at near-coastal cities such as Vancouver, BC; Seattle, WA; and Portland, OR (Figure 1). This paper summarizes and analyzes the constraints on this downdip limit. It focuses on Cascadia, but many of the limits are applicable to the seismogenic zones of other subduction zones. It also deals principally with the main part of the 800 km long rupture zone between central Vancouver Island south of the Nootka Fault zone and southern Oregon, which generates ~M9 earthquakes. The more complex areas to the north and south may be involved in the long M ~ 9 ruptures but are also large enough to generate separate events up to M8. The areas of the Explorer plate off northern Vancouver Island and the Gorda region off Northern California are subducting younger ocean lithosphere than the main 1 Pacific Geoscience Centre, Geological Survey of Canada, Sidney, British Columbia, Canada. 2 School of Earth and Ocean Sciences, University of Victoria, Victoria, British Columbia, Canada. Corresponding author: R. D. Hyndman, Pacific Geoscience Centre, Geological Survey of Canada, 9860 W. Saanich Rd., Sidney, BC V8L4B2 Canada. ([email protected]) ©2013. American Geophysical Union. All Rights Reserved. 2169-9313/13/10.1002/jgrb.50390 central region and may have a shallower downdip limit because of the higher inferred temperatures. The main Cascadia rupture appears to occur primarily in M9 events that rupture much of if not the whole length of the margin based mainly on the deep-sea record of earthquake-generated turbidites and the coastal marsh coseismic subsidence [e.g., Goldfinger et al., 2012; Leonard et al., 2010] and with a generally consistent downdip limit [e.g., Leonard et al., 2010]. This is in contrast to many other subduction zones that exhibit a wide range of rupture areas and earthquake magnitudes. The model of Flück et al. [1997] for the thrust locked zone based on geodetic data has been used as a reference for comparison among the various constraints to the downdip seismogenic limit (Figure 2). As discussed below, thermal limits to downdip seismic behavior are similar. [3] Understanding the downdip limit provides important information on what controls the areas of rupture on subduction thrust faults globally and the nature of fault movement by seismic and aseismic motions. Also critically important to this discussion are well-recorded recent M ~ 9 great earthquakes that provide method calibration, i.e., Tohoku Japan 2011, Chile 2010, and Sumatra 2004, although these great earthquakes occurred where older colder ocean lithosphere is being subducted such that the seismogenic limit may not be thermally controlled, in contrast to the inferred thermal control for Cascadia. Relevant information comes from geological study of deeply exposed thrust faults which show that subduction thrusts may be very complex shear zones on 5530 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Figure 1. Schematic of the downdip limit of Cascadia great earthquake rupture, the closest approach to near-coastal cities. Episodic tremor and slip (ETS) is well landward of the rupture limit and approximately corresponds to the fore-arc mantle corner. a geological outcrop scale with an active layer of underthrust sediment in a subduction melange [e.g., Bachmann et al., 2009; Fagereng, 2011]. Thrust structure complexity with a thick duplex structure, mainly deeper downdip than the seismogenic zone for northern Cascadia has been proposed by Calvert et al. [2011] using multichannel seismic reflection data. Some of the constraints have been discussed in previous reviews of the seismogenic zone of subduction faults [e.g., Tichelaar and Ruff, 1993; Pacheco et al., 1993; Oleskevich et al., 1999; Hyndman, 2007; Hyndman and Rogers, 2010; Heuret et al., 2011]. [4] The updip limit of rupture toward the trench also is important for estimating earthquake magnitude and especially for modeling tsunami generation. For many other subduction zones, it is often concluded that the updip rupture limit does not reach the trench (e.g., summary by Hyndman et al. [1997]). The 2011 M9 Tohoku rupture did reach the trench with large displacement [e.g., Fujiwara et al., 2011] but at such a slow rupture speed that there was little energy in the seismic frequencies. However, the very large tsunami generation appears to have been generated mainly in the updip region. Bilek and Lay [1999] argued that the rigidity is very low and therefore the rupture speed is very low at the toe of accretionary sedimentary prisms, increasing downdip perhaps because of clay consolidation. If the updip seismogenic limit is at the depth where stable-sliding clays dehydrate, i.e., 100–150°C [e.g., Moore and Saffer, 2001; Oleskevich et al., 1999], this limit may be near the deformation front for the high-temperature Cascadia subduction zone. However, laboratory data by Marone and Saffer [2007] suggest that other physical changes with depth must be responsible for the updip part of the thrust commonly being aseismic. The Cascadia updip limit is poorly constrained and is not discussed further in this paper. [5] The downdip seismogenic limit is where the fault zone no longer exhibits frictional instability [e.g., Scholtz, 2002] (Figure 1). Many factors have been suggested for the controls of the downdip landward extent (e.g., discussions by Tichelaar and Ruff [1993], Pacheco et al. [1993], Oleskevich et al. [1999], and Hyndman [2007]), but two have been argued to be most important. For hot subduction zones where young lithosphere is being underthrust like Cascadia, there appears to be a maximum temperature limit for seismic behavior [e.g., Hyndman and Wang, 1993; Oleskevich et al., 1999]. For the more common cool subduction zones, aseismic serpentinite and talc downdip in the fore-arc mantle corner may limit the downdip rupture [e.g., Ruff and Tichelaar, 1996; Hyndman et al., 1997; Oleskevich et al., 1999; Peacock and Hyndman, 1999; Hyndman and Peacock, 2003]. Ruff and Tichelaar [1996] noted that the downdip limit is often near the coast. Other factors that could affect the downdip rupture limit are variations in the stress regime of the thrust fault zone, pore pressure on the fault, fault roughness, and upper and lower plate structure [e.g., Scholz, 1992; Pacheco et al., 1993; McCaffrey et al., 2008; Llenos and McGuire, 2007; Heuret et al., 2011; Bilek, 2007; Wallace et al., 2009; Wang and Bilek, 2011]. [6] The location of the seismic source and interpretation of a number of the downdip constraints depend on the dip angle Figure 2. Fully locked (to solid black line) and linear transition (to dashed gray line) zones based on leveling and tide gauge geodetic data, used as a reference for comparison among the various landward seismic limit constraints (modified after Flück et al. [1997]). The thermal limits of seismic behavior, 350 and 450°C, from thermal models are shown with uncertainties, as given by Hyndman and Wang [1995]. The limits from this thermal model and geodetic limits are in agreement within the uncertainties. ETS is well landward. 5531 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Figure 3. Schematic geometry of the Cascadia subduction thrust, showing the simple model of locked or “full rupture” and transition zones (adapted from Flück et al. [1997]). The shallow thrust angle for northern Washington results in the temperature limits being farther landward and wider locked/ rupture and transition zones. and profile of the subducting slab (Figure 3) [e.g., Flück et al., 1997; McCrory et al., 2012a]. The principal large-scale variability is the region of shallow dip beneath the margin of northern Washington and southern Vancouver Island. As discussed below, it appears that the downdip rupture limit may be thermally controlled. Therefore, the region with shallow dip is expected to reach a critical temperature farther landward and have a wider rupture zone than the steeper dip regions to the south and north. Most but not all past great events appear to have ruptured all of the main part of the Cascadia subduction fault [e.g., Goldfinger, 2011; Goldfinger et al., 2012; Leonard et al., 2004; 2010] but there have been some that ruptured shorter extents along the subduction fault. The distribution of event sizes and margin-parallel extents is the subject of ongoing research [e.g., Atwater and Griggs, 2012; Frankel, 2011]. [7] We discuss nine constraints to the landward limit of Cascadia great earthquake rupture. These nine constraints provide a better understanding of the landward seismogenic downdip limit and the controlling processes than are available for many other subduction zones, even those which have had instrumentally recorded great events. Each of the nine constraints requires different assumptions and has different uncertainties; we seek a rupture limit that is consistent with all of the constraints. There is a considerable literature on aspects of some of these constraints. An extensive reference list is given, but far from complete. Emphasis is on those papers that provide general results and those that have reference lists themselves that lead to papers with more detailed information. [8] An important issue is the nature of the landward tapering of rupture displacement, i.e., the “transition zone” (Figure 3). This transition is important since 1–2 m of displacement can produce strong ground motions in the main damaging frequency range, as now well recognized, for example, the recent M9 Tohoku great earthquake off NE Japan [e.g., Kurahashi and Irikura, 2011; Koper et al., 2011] and for the M9 2004 Sumatra and 2010 Chile earthquakes and a number of other subduction zones [e.g., Lay et al., 2012] where much of the energy in the damaging frequencies was from the downdip part of the rupture. For some constraints, it is useful to define the area of at least 50% or ~10 m of the estimated ~20 m average full rupture (see Figure 4). The position of 10% or 2 m, which in the geodetic dislocation models is concluded to be tens of kilometers farther landward, is much less constrained than 50% because the former depends on the nature and extent of the transition from full to zero rupture displacement. Also, the nature of the actual transition is likely quite variable among different events. Geodetic locked zone rupture inversions usually give smooth peak rupture and landward decrease. [9] The degree to which the downdip limit to rupture varies among individual events remains uncertain. Most Cascadia events appear to rupture a reasonably consistent landward extent based on paleoseismicity coastal marsh subsidence data [Leonard et al., 2010]. Although there undoubtedly is variability, the summarized estimates are taken to be the maximum landward rupture extent for most events, as to be expected if there is a physical control of this limit that does not change with time. [10] There are three types of constraints: (a) constraints to the currently “locked zone” on the thrust that is building elastic strain that will be released in future great earthquakes (constraint 1 below), (b) constraints to the landward extent of past great earthquake rupture (constraint 2 below), and (c) physical process/state controls on where seismicity can and cannot occur on the thrust (constraints 3–9 below). It is important to recognize that none of the constraints directly constrain the actual rupture limit; all require some assumptions. [11] The constraints are: [12] 1. The “locked/transition” zones on the thrust from modeling geodetic observations of current deformation (GPS, repeated leveling, tide gauges, etc.). [13] 2. Past great earthquake rupture zone constrained by paleoseismic coastal marsh subsidence, 1700 and earlier events, i.e., “paleogeodesy”. [14] 3. Seismic behavior limits from downdip temperatures, loosely described as the “brittle-ductile transition”, approximately 350°C for where the rupture can initiate and there is full rupture, 450°C for the landward limit of the transition zone from full rupture to zero displacement. Figure 4. Locked and transition zone models corresponding approximately to the full rupture, and linear and exponential tapering rupture (effective transition zone; adapted from Wang et al. [2003]). The horizontal distance scale and rupture magnitudes are approximately those for Vancouver Island and northern Oregon. 5532 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT 2. Locked/Transition Zones From Geodetic Deformation Figure 5. Schematic of margin deformation through the great earthquake cycle illustrating the interseismic uplift and coseismic subsidence of the coastal region. The coseismic subsidence for most great events, up to 2 m, is about 500 years times the uplift rate per year. Most repeated leveling and other vertical geodetic data are landward of the peak uplift and most coastal paleoseismic subsidence data are near the maximum subsidence. [15] 4. Change in seismic reflection character from a thrust marked by a thin sharp reflection to a thrust zone marked by a thick set of reflectors, i.e., brittle to ductile shear deformation. [16] 5. Fore-arc mantle corner, landward of which there is inferred aseismic serpentinite and talc overlying the thrust. [17] 6. Updip limit of interseismic episodic tremor and slip (ETS) slow slip that accommodates most of the long-term plate convergence. [18] 7. Empirical association of rupture area with shelf-slope sedimentary basins defined mainly by gravity anomalies. [19] 8. Landward downdip limit of small thrust earthquakes on the subduction thrust that indicate seismic behavior. [20] 9. Landward limit of earthquakes on the Nootka transform fault zone as it subducts beneath the margin of Vancouver Island. [21] Very general downdip rupture limits are provided by estimates of tsunami wave amplitudes and runups, although the updip extent and displacement are a more important factor. Only the 1700 tsunami has been at all well documented for Cascadia so this constraint is not discussed here. We have used the Flück et al. [1997] model with a linear downdip transition zone as a simple reference for comparison among the different constraints. Other models with a linear transition zone give similar results [e.g., Mazzotti et al., 2007], as do most inversion models. [22] The most studied and one of the strongest constraints to the rupture zone is the definition of the locked zone on the thrust fault through “back-slip” modeling [e.g., Savage, 1983] of geodetic data (GPS, repeated leveling, tide gauges, etc.). This approach is based on the approximation that the earthquake cycle is elastic to a first order, i.e., elastic strain builds up nearly linearly with time since the previous event and is released completely by megathrust earthquakes (Figure 5). The pattern of surface deformation for the rupture is then close to the mirror image of the interseismic rate times the earthquake interval (average 500–600 years for the main Cascadia events). There are two significant limitations. First, there may not be full “coupling”; there may be some aseismic slip between great earthquakes (i.e., discussion by Wang and Dixon [2004]). The current geodetic data that are consistent with a fully locked fault [e.g., Hyndman and Wang, 1995; McCaffrey et al., 2013] and the lack of significant interseismic thrust seismicity suggest that there is currently little interseismic slip. The thrust is fully locked after some postseismic transients. Also, the estimated coseismic rupture displacement in past events approximately equals the interseismic plate convergence slip deficit based on the magnitude of coseismic coastal subsidence [Leonard et al., 2004; 2010]. The one area of uncertainty is off central Oregon where the inferred locked/rupture zone is well offshore such that the expected and observed interseismic uplift rate and coseismic subsidence are small. From these two types of data, it is therefore difficult to exclude some degree of interseismic aseismic creep motion for the thrust in this region with the partially seismic zone extending farther landward [e.g., Pollitz et al., 2010; Wang, 2012]. This uncertainty is discussed below. [23] Second, because the Earth is viscoelastic, there are theoretical problems of uncertain magnitude in the elastic backslip model for the locked zone, especially postseismic transients [e.g., Hu et al., 2004; Wang 2007; Kanda and Simons, 2010; Hetland and Simons, 2010; Suito and Freymueller, 2009; Wang et al., 2012]. The downdip extent of fully seismic rupture is expected to be somewhat seaward of the limit for the modeled interseismic locked zone because some release of the locked zone strain buildup, especially the deep part, is in postseismic slip [e.g., Wang, 2007; Hu and Wang, 2012; Wang et al., 2012], so the limit from the geodetic data may be slightly landward of the actual coseismic full rupture limit. Records of local deformation as a function of time for a few subduction zones that have been adequately monitored indicate that the long-term interseismic strain buildup is approximately linear after a sometimes substantial postseismic transient that may depend on the earthquake magnitude [e.g., Wang, 2007]. However, this question requires more rigorous analysis and calibration by recent great earthquakes elsewhere. An example from the M ~ 8 1944 and 1946 events off southwest Japan is shown in Figure 6 that indicates good correspondence of the preseismic pattern of deformation with that for the coseismic rupture deformation. The downdip rupture extent of defined by dislocation modeling of leveling data was very close to that for the locked zone defined by prior geodetic data, within about ±10 km [Hyndman et al., 1995; see also Sagiya and Thatcher, 1999]. This agreement suggests that the postseismic transient effect 5533 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Figure 6. SW Japan 1944/1946 earthquakes, comparison of coseismic vertical displacements and interseismic vertical motion from repeated leveling measurement, and dislocation models with a linear transition zone (modified from Hyndman et al. [1995]). The coseismic vertical has been inverted and scaled by the average earthquake cycle duration. The horizontal lines at the top show the three widths of the locked (thick line), transition (thin line), and free slip zones for the three vertical models shown, i.e., ±10 km from best fit model (blue curve). The good agreement between the interseismic locked zone and the coseismic rupture zone deformations indicates that the geodetically determined locked/transition zones are good estimators for the rupture/transition zone. is quite small for at least these megathrust events. There also is good correspondence of the modeled locked zone using geodetic data and the main downdip rupture zone for the 2010 M8.8 Chile event [e.g., Delouis et al., 2010; see also Ruegg et al., 2009], the 2004 M9 Sumatra earthquake [e.g., Shearer and Burgmann, 2010, and references therein], and the 2011 M9 Tohoku, Japan event [e.g., Loveless and Meade, 2011]. The interseismic locked zone estimated a few decades after the last great earthquake is likely a reasonable estimator of the rupture zone for Cascadia, as discussed for past great earthquakes off SW Japan. [24] Another possible effect on the use of the locked/transition zone as a rupture estimate is that the zone of ETS slow slip downward of the main rupture zone is locked or nearly locked on a timescale averaging about 14 months [e.g., Rogers and Dragert, 2003]. Strain on this short-term locked zone is released during a period of several weeks of slow slip. However, if longer-term average geodetic data are used, the effect of the slow slip zone events on the estimate of the main locked zone and the estimate of the subsequent rupture zone is small [e.g., Holtkamp and Brudzinski, 2010]. [25] Both forward models and inversions have been applied to the backslip modeling with similar results. The inversion results generally have a longer taper inland as a consequence of the smoothing required in the method, unless the transition taper width is constrained in the inversion [e.g., McCaffrey and Schmidt, 2012; McCaffrey et al., 2013]. Most recent models used profiles with smoothly increasing dip landward [e.g., Hyndman and Wang, 1995], and Flück et al. [1997] developed a 3-D fault model that has been updated by McCrory et al. [2006; 2012a] and used in most subsequent modeling. A simple fully locked zone and linear transition zone have been used in the initial models, but a more appropriate exponential transition zone was proposed by Wang et al. [2003] (Figure 4). [26] Vertical geodetic data give a quite well-defined limit to rupture because both coseismic and interseismic vertical displacements change from upward to downward approximately at the limit of significant rupture on the fault, i.e., hinge line (Figures 5 and 6). Most Cascadia vertical constraints have come from repeated high-precision leveling and long-duration tide gauges [e.g., Mitchell et al., 1994; Hyndman and Wang, 1995; Verdonck, 2005; Burgette et al., 2009]. GPS vertical data have lower resolution than horizontal data and are just beginning to have useful accuracy of a few mm/yr [e.g., Mazzotti et al., 2007; McCaffrey and Schmidt, 2012]. Highresolution gravity and borehole strainmeters also are providing additional strain data for comparison with dislocation models [e.g., Mazzotti et al., 2007; Henton et al., 2010]. An example of vertical releveling data and models depicting the observed and predicted pattern of coastal uplift are shown in Figure 7 [Hyndman and Wang, 1995; Burgette et al., 2009]. Note that Figure 7. Examples of leveling data and corresponding dislocation models across the Cascadia margin (modified from Hyndman and Wang [1995]). The rates of vertical motion derived from leveling data are compared to those predicted by the three different width dislocation models shown; the labels give the model locked and transition zones widths in kilometers and the average dip angle for the locked part of the thrust. Detailed data for the Oregon margin have been presented by Burgette et al. [2009]. 5534 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Figure 8. (a) Cascadia GPS vectors of current deformation (modified from McCaffrey et al. [2007]). The motion rates include the effect of elastic strain buildup on the subduction thrust and fore-arc crustal motions. (b) The elastic dislocation model from inversion of horizontal and vertical data. The red dashed line is approximately 50% locking (~25 mm/yr) (modified from McCaffrey et al. [2013]). A locked or nearly fully locked thrust and transition zone is indicated for the whole margin, very similar to the Flück et al. [1997] reference model shown as black solid and dashed lines. the maximum uplift rates of about 1– 4 mm/yr at the coast from geodetic data, multiplied by the average interevent period of 500–600 years, predict maximum coseismic subsidences of 0.5–2 m, in general agreement with those observed in the coastal marsh data along the coast [e.g., Leonard et al., 2010]. The patterns of the locked and transition zones for all of the studies are very similar to those of the Flück et al. [1997] reference model, with successive improvements from more accurate data and dislocation parameters. Burgette et al. [2009] inferred an offset to a shallower locked depth over the southern end of the coastal Siletz Terrane, but the effect is small. [27] Horizontal GPS data show a clear signal of the coastal region being squeezed landward [e.g., McCaffrey et al., 2013, Figure 8a]. The vectors also show the strong forearc crustal deformation that must be accounted for when conducting megathrust dislocation modeling. Forward dislocation modeling that matches observations of landward shortening has been presented by many authors with successive improvements from more extensive and accurate data and dislocation parameters [e.g., Dragert and Hyndman, 1995; Flück et al., 1997; Miller et al., 2001; Wang et al., 2003; Mazzotti et al., 2003, 2007]. Wang et al. [2003] used an effective transition zone that decreased exponentially landward to allow for the postseismic transients (Figure 4). The recent models corrected for the effect of the Oregon block fore-arc block motion, i.e., used the convergence across the near-coastal margin rather than the ocean plate relative to North America. [28] Backslip inversions of the GPS and vertical data have been presented by Verdonck [2005], Yoshioka et al. [2005], and McCaffrey et al. [2007, 2013] which are generally in agreement with the results of forward modeling. In the most recent inversion analysis McCaffrey et al., [2013] found a good fit of both horizontal and vertical data (Figure 8b). They concluded that there is a somewhat wider transition zone off central Oregon compared to southern Oregon and northernmost California but the differences compared to previous models are quite small. [29] From modeling both horizontal and vertical data, the estimated full rupture (locked zone) is offshore everywhere except near Cape Mendocino in Northern California, with model linear transition zones extending onshore in northern Washington and near the coast in Oregon and Vancouver Island. The recent inversion model from McCaffrey et al. [2013] is shown in Figure 8b. This 50% locked location is very close to that for the Flück et al. [1997] reference model, i.e., middle of the transition zone. It is not yet clear how much of the small-scale variability from this type of inversion is real or due to data uncertainty. Smooth inversion transition zones have rupture downdip displacement of 2 m or ~10% of full rupture near the landward limit of the linear forward model transition zones, but this part of the models is very poorly constrained. 3. Rupture Zone From Paleoseismic Coastal Marsh Subsidence in Great Earthquakes [30] The abrupt vertical deformation associated with the most recent great earthquake in 1700 and earlier events recorded by subsidence in coastal intertidal marshes (Figure 9) can be used to constrain the area of rupture, i.e., “paleogeodesy” [Leonard et al., 2004; 2010; Wang, 2012; Hawkes et al., 2011; Peterson et al., 2012]. Deeper-buried marsh surfaces are covered by sediments deposited after successive coseismic subsidence events of up to 2 m (e.g., Figure 9a). Although the coseismic subsidence is approximately recovered through strain buildup in the interseismic period, ongoing sea level rise results in progressive burial of subsided marshes [e.g., Atwater, 1987]. On the coast of Vancouver Island, there is ongoing land uplift faster than sea level rise so only the 1700 buried marsh remains [Clague and Bobrowsky, 1994]. The vertical distance between the present marsh top and the 1700 marsh is not the coseismic subsidence because of eustatic sea level rise, the elastic strain buildup since that event, and tectonic vertical motions. However, a rough estimate can be obtained from this vertical distance after correcting for these effects. Better displacement estimates come from careful calibration of the elevation range of a variety of coastal organisms, allowing accurate estimation of the paleo-elevation differences between the sediments just above and just below the top of the buried marshes. The resolution is commonly ~0.5 m but may be as good as ±0.3 m [e.g., Guilbault et al., 1996; Peterson et al., 2012; Hawkes et al., 2011]. 5535 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Figure 9. (a) Example of a buried intertidal marsh off central Washington attributed to coseismic subsidence during the M9 earthquake of 1700 (photo by L. Leonard). A tsunami-driven sand sheet overlies the buried marsh. (b) Comparison of coastal coseismic subsidence estimates for an area of the Oregon coast plotted with distance orthogonal to the coastal trend, and corresponding dislocation models (location shown in Figure 10), for the 1700 and pre-1700 events [after Leonard et al., 2010]. The model results are shown for a 14 m full downdip rupture displacement and for the rupture that releases the total shortening elastic strain for a 500 year return period, using the variable convergence rates along the margin. [31] For individual events, there is some trade-off between the amount of rupture displacement and the landward limit of rupture, but the change from seaward coseismic uplift to landward coseismic subsidence, i.e., hinge line, is approximately at the full displacement rupture limit (see Figures 5 and 6). Because coseismic coastal subsidence is observed for most of the Cascadia margin, this hinge line and the full coseismic rupture are concluded to be well offshore. The only exception is for a small area just north of Cape Mendocino where there has been coastal coseismic uplift and full rupture displacement is inferred to extend to beneath the coast. This hinge line constraint limits the uncertainty in the landward limit from this trade-off. Where there is only a very small coastal subsidence in central Oregon, there is an additional uncertainty [e.g., Wang, 2012; Pollitz et al., 2010]. [32] The observed coastal subsidence for the 1700 event and the average of previous events compared to the predictions of the backslip model from Leonard et al., [2010] for the whole margin are shown in Figure 10a along with the high-resolution coastal subsidence data from Hawkes et al. [2011] for Oregon shown in Figure 10b. There is good agreement of the observed coastal subsidence rupture limits with the estimates from the geodetic data locked and transition zones, within the considerable downdip uncertainties of 20–40 km. Wang [2012] modeled the coseismic subsidence pattern with rupture that varied along the margin to get a better fit to a section of the coast with high-resolution data. A slightly better model fit was found to the data but the misfit differences compared to those for rupture scaled directly to convergence rate along the margin are within the uncertainties. 4. Downdip Temperatures: The Seismic-Aseismic Transition [33] Above a critical temperature, rocks deform by ductiletype mechanisms and seismic rupture is not expected, as much discussed for continental transcurrent faults like the San Andreas [e.g., Chen and Molnar, 1983; Wong and Chapman, 1990; Ito, 1990]. The transition from brittle to ductile behavior is closely related but not identical to the limit of seismic instability [e.g., Scholz, 1998]. For seismic behavior, the critical change downdip is from velocity-weakening to velocity-strengthening slip. With velocity weakening, the slip surface weakens with ongoing displacement and the effective coefficient of friction decreases. Runaway earthquake displacement can occur that releases much if not all of the built-up elastic energy. For velocity strengthening, the slip zone strengthens with ongoing displacement so slip speeds are limited, a viscous-like behavior. For felsic crustal rock compositions appropriate for the continental crust overlying the thrust, for accretionary sedimentary prisms, and subduction channel sediments, the velocity-weakening to velocitystrengthening transition from laboratory data is about 350°C [Tse and Rice, 1986; Blanpied et al., 1991; 1995; Bürgmann and Dresen, 2008] which agrees reasonably well with the estimated temperatures at the maximum depth of seismicity in continental fault zones [e.g., Chen and Molnar, 1983; Wong and Chapman, 1990; Ito, 1990; Bonner et al., 2003]. The mafic subducting oceanic crust may have a higher seismogenic temperature limit, but the thrust deformation is expected to occur in the weaker overlying sediments or forearc crust. The exception of local areas where there is shearing 5536 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Figure 10. (a) Along the coast comparison of the magnitude of coastal marsh subsidence and predictions from dislocation models for 200 and 500 years of strain buildup at the convergence rates shown [after Leonard et al., 2010]. The best fit is for approximately a 500 year great earthquake return period indicating that most of the interseismic plate convergence is taken up as elastic strain. The ellipse shows the location of the data in Figure 9. (b) High-resolution 1700 coseismic subsidence estimates for Oregon compared to elastic dislocation models (modified from Hawkes et al. [2011]). The shaded vertical bar has lower accuracy. The gray band is the mean estimate of coastal subsidence from Leonard et al. [2010]. of mafic seamounts, etc., having a higher temperature limit, is discussed below. Also, as discussed below, aseismic serpentinite/talc may be important downdip of the fore-arc mantle corner. [34] At the landward downdip end of great earthquake rupture, there must be a transition zone within which rupture cannot initiate but into which slip can continue downdip if rupture is initiated updip. This zone represents “conditional stability” [e.g., Scholz, 2002]. The downdip limit of the transition zone is expected to be where the instantaneous shear stress for slip increases rapidly downward, for example, from the effect of increasing temperature. Laboratory studies suggest this limit is at about 450°C for crustal rocks [e.g., Blanpied et al., 1991; 1995]. Thermal model estimates of downdip temperatures are then required to define the location of 350°C and 450°C on the thrust. [35] Recent subduction thermal models use finite element modeling, a landward steepening thrust profile, accurate incoming plate thermal regimes that include the effect of overlying insulating sediments, and the effect of radioactive heat generation and of variable thermal conductivity, e.g., Figures 11a and 11b [Hyndman and Wang, 1993, 1995; Hyndman et al., 1995]. Their map locations of the 350 and 450°C isotherms on the thrust are shown in Figure 2. The locations are within the uncertainties of the limits of the reference model locked and transition zones from geodetic data. It has been argued that the effect of frictional heating on the thrust in the depth range of the seismogenic zone is small and usually may be neglected [Wang et al., 1995; Wang et al., 1995; Wang and He, 1999] but the magnitude of heating remains a source of uncertainty [see McKenna and Blackwell, 2002]. Currie et al. [2004] included convection in the back arc in the models but this made no significant difference to temperatures in the region of the seismogenic zone. Gutscher and Peacock [2003] proposed a flat slab profile for the Cascadia subducting slab and estimated about 30 km farther landward positions of the 350 and 450°C temperatures on the thrust compared to the previous models. However, the compilation of McCrory et al. [2006, 2012a] with more recent data does not show the degree of slab flattening used in their Cascadia models. [36] Cooling of the crust beneath the margin by hydrothermal circulation has been proposed by Cozzens and Spinelli [2012] such that the critical thermal limits could be as much as 30–55 km landward relative to results from simulations without fluid flow. Heat may be transferred from beneath the accretionary sedimentary prism to the adjacent deep-sea basin through fluid flow in the subducting crust, decreasing the heat flow beneath the continental slope and increasing the heat flow from the adjacent deep sea floor. Off SW Japan where the incoming sediment section is much thinner, the effect of hydrothermal circulation on the thermal data is clear [Spinelli and Wang, 2008], with heat flow seaward of the deformation front much higher than for a conductive model. However, off Cascadia where the sediment section offshore is very thick and there may be little crustal hydrothermal circulation, there is no indication of high heat flow seaward of the deformation front; the heat flow corresponds to that expected for a cooling plate model after correction for the sedimentation effect. Landward of the deformation front, the best model fit to the heat flow data across the Vancouver Island margin is for little or no hydrothermal circulation cooling effect (Figure 11). On this profile, the heat flow has been corrected for the thermal effect of accretionary prism sediment thickening [Hyndman et al., 1993]. The ~15% effect on the thermal model is shown in Figure 11a. It is likely that the models with little or no effect from hydrothermal circulation would also agree with the thermal data off Washington, Oregon, and California if a correction was made for the sediment thickening reduction effect. However, Cozzens and Spinelli [2012] also argued that the effect may be important based on the depth to the basalt-eclogite transition indicated in receiver function structure data, which emphasizes the uncertainties in the thermal model predicted thrust temperatures. 5. Change in Thrust Seismic Ref lection Character From Thin Sharp Seismic to Thick Aseismic Ductile Shear Zone [37] From the thermal analysis summarized above, it is expected that there is a thermally controlled downdip change from brittle-seismic to ductile-aseismic behavior at about 350°C–450°C. This transition roughly corresponds to the transition as seen in field geology studies from cold brittle type behavior and thin fault zones, to broad mylonitic, and then ductile shear zones [e.g., Sibson, 1977, 1983; Cole, et al., 2007]. The thin shear represents strain weakening 5537 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Figure 11. (a) Heat flow data across the northern Cascadia margin compared to the thermal model prediction of Hyndman and Wang [1995]; the dashed red line shows the prism thickening effect that depresses the surface heat flow. (b) Thermal cross section showing the thermal downdip limits to postulated locked (350°C) and transition (450°C) zones. The 350 and 450°C isotherms and uncertainties are shown on the map of Figure 2, compared to the locked and transition zones from modeling geodetic data. If ocean crustal hydrothermal cooling is important [Cozzens and Spinelli, 2012], these temperature limits may be several tens of kilometers landward. whereas the thick shear likely represents strain strengthening. In the latter, the fault zone continually moves outward from the initial shear zone which becomes stronger with deformation [e.g., Kirby, 1985] and thickens into weaker as yet undeformed rocks. The transition from thin brittle to thick ductile behavior is expected to be at a slightly higher temperature than the limit of seismic instability [e.g., Scholz, 2002] and the comparison needs better calibration. [38] As described by Nedimovic et al. [2003], the change from a thin strong detachment reflector to a reflective 5–10 km thick band in marine and land multichannel seismic reflection data approximately corresponds to the downdip limit of seismic rupture for M > 8 earthquakes at a number of subduction zones, e.g., Alaska, Chile, and SW Japan. They concluded that the thin to thick reflection transition marks the brittle to ductile transition. They mapped this transition along the margin of southern Vancouver Island (reflection zone 2 on map of Figure 12) and found it generally lies just seaward of the coast and corresponds reasonably well to the modeled transition zone downdip limit from other constraints, i.e., approximately 450°C. Despite considerable variation in the reflection character of the subduction thrust on a number of seismic lines, there is a consistent downdip change from a thin reflection zone offshore (reflection zone 1) to a broad reflection band (E layer; reflection zone 3) some 5 to 7 km thick, near the west coast of Vancouver Island (Figure 12). The thin reflection is less than 2 km thick, but the true boundary is probably thinner, the image being thickened by the reflection signature. The transition is a few tens of kilometers wide that corresponds in position very approximately to the downdip limit of the geodetic and thermal transition (450°C) zones (Figure 12). [39] This interpretation across the Vancouver Island margin assumes that the reflective E-zone [e.g., Clowes et al., 1987] is the detachment. However, other authors [e.g., Audet et al., 2009; Hansen et al., 2012] preferred the interpretation that the reflective band is the subducting oceanic crust or the upper crust. In contrast, from detailed seismic structure analysis, Calvert et al. [2011] concluded that the E-layer dipping reflective band is a shear zone above the subducting plate as in the original interpretation of Clowes et al. [1987]. This disagreement is not yet resolved. Further calibration of this method using the change in reflection character downdip is needed for subduction zones that have had recent great earthquakes with a well-constrained downdip rupture limits. Initial results from a seismic structure survey to examine the downdip change in reflection character for the Alaska subduction zone were reported by Shillington et al. [2011]. 6. Fore-Arc Mantle Corner: Aseismic Serpentinite and Talc on Thrust [40] The fore-arc mantle corner may limit downdip rupture because of the presence of aseismic serpentinite and talc. The fore-arc mantle is commonly significantly serpentinized by rising fluids driven upward from the underlying dehydrating subducting crust [e.g., Hyndman and Peacock, 2003, and references therein] (Figure 13). The structure of serpentinite is expected to make it weak where it is in contact with the underlying fault zone, a conclusion supported by the arguments for no significant frictional heating on the thrust, at least landward of the fore-arc mantle corner [e.g. Wada et al., 2007]. There is some uncertainty as to the weakness 5538 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Figure 12. Change in megathrust multichannel seismic reflection character from a thin inferred seismic signature (Reflection band 1), to a thick inferred aseismic shear zone (Reflection band 3; modified after Nedimovic et al. [2003]), compared to Flück et al. [1997] reference locked and transition zones from modeling geodetic data, i.e., approximately 350°C and 450°C. The reflection transition (Reflection band 2) occurs at slightly cooler than 450°C and seaward of the ETS zone. of serpentinite [e.g., Escartın et al., 2001; Moore et al., 1997; Moore and Rymer, 2007; Reynard, 2013], but another factor may ensure that the thrust beneath the fore-arc mantle is usually aseismic. Talc is generated by the silica-saturated fluids from the underlying dehydrating oceanic crust and underthrust sediments that react with the overlying fore-arc mantle serpentinite [e.g., Peacock and Hyndman, 1999]. Talc is a very weak mineral and fault zones containing talc are very likely aseismic, as suggested for the San Andreas Fault [Moore and Rymer, 2007]. Brocher et al. [2003], Preston et al. [2003], and Ramachandran and Hyndman [2011] using wide-angle seismic structure and seismic tomography studies showed that there is a substantial reduction in seismic velocity and increase in Poisson’s ratio in the Cascadia fore-arc mantle, indicative of extensive serpentinization, and Blakely et al. [2005] showed that the fore-arc mantle serpentinization produces a significant magnetic anomaly high. [41] Tichelaar and Ruff [1993], Hyndman et al. [1997], and Oleskevich et al. [1999] concluded that the maximum depth of great earthquake rupture, commonly defined by aftershocks, often corresponds to the intersection with the fore-arc mantle corner. Unfortunately, the location of the fore-arc mantle corner has not yet been well defined on most margins. Ruff and Tichelaar [1996] concluded that for the South American subduction zone, the downdip limit of great earthquake rupture approximately coincides with the fore-arc mantle corner which in turn is approximately at the coast for much of that margin. Heuret et al. [2011] compiled the maximum depth of intermediate magnitude thrust events for a number of M5.5–7.0 thrust earthquakes in a global study and concluded that the average maximum event depth for subduction zones where ocean plates descend beneath continental lithosphere was about 40 km. This is somewhat deeper than the average of 34 ± 11 km for the compiled depth estimates of fore-arc Moho depth by Wada and Wang [2009]. However, this is not a large difference since the Moho intersection depth for most subduction zones is very poorly defined. For example, the Moho may not exhibit a crust-mantle downward velocity increase because of the low-velocity fore-arc mantle serpentinite [e.g., Bostock et al., 2002]. Therefore, this conclusion remains uncertain. [42] Three recent well-studied great earthquakes allow examination of this proposed downdip rupture limit for M ~ 9 events, Sumatra in 2004, Chile in 2008, and NE Japan in 2011. For the great earthquake of 2004 off Sumatra, the rupture extent was to a depth of about 30 km [e.g., Chlieh et al., 2007, and references therein] in agreement with the locked zone from geodetic and paleogeodetic data [e.g., Chlieh et al., 2008]. The depth of ~30 km is within the common range of 30–50 km for continental fore-arc mantle corners, but Dessa et al. [2009] and Simoes et al. [2004] have concluded from seismic structure data that the fore-arc Moho depth is very shallow, 21–25 km, on this margin such that the locked/rupture zone extends to beneath the fore-arc mantle. If this unusually shallow Moho depth does apply to this fore-arc mantle corner, the corner does not constrain downdip rupture on this margin. However, the inference of an exceptionally shallow fore-arc mantle leaves considerable uncertainty. Also, the depth and the critical temperature for seismic behavior are similar [Hippchen and Hyndman, 2008], so the downdip limit may be thermally controlled rather than by the fore-arc mantle corner. [43] For the great earthquake off Chile in 2008, Tong et al. [2010] from InSAR data and Delouis et al. [2010] from geodetic data concluded a downdip coseismic rupture limit near the depth where the subducting Nazca plate intersects with the continental Moho of the South America plate at 40–50 km [Haberland et al., 2006]. This depth is in the range of maximum rupture depths estimates for a compilation of earlier large thrust earthquakes on this margin by Tichelaar and Ruff [1991]. This depth also agrees with the limit of the locked zone inferred from preseismic GPS data [Ruegg et al., 2009; Moreno et al., 2010]. Haberland et al. [2006] found the fore-arc Moho at a depth of about 50 km in good agreement with the fore-arc mantle corner being the downdip limit for rupture in this great earthquake. [44] For the third recent great earthquake, Tohoku off NE Japan in 2011, the main large displacement rupture appears 5539 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Figure 13. Cascadia section showing the inferred aseismic serpentinite and talc on the subduction thrust that may limit seismic rupture downdip (modified from Peacock and Hyndman [1999]; bru = brucite; serp = serpentinite). The ETS zone approximately corresponds to the fore-arc mantle corner, whereas the estimated Cascadia seismogenic zone is significantly shallower. to terminate just seaward of the coast where the thrust depth is 30–40 km deep [e.g., Simons et al., 2011; Loveless and Meade, 2011], a common depth for the fore-arc mantle corner. Aftershocks extend to a depth of 40–45 km just seaward of the coast. However, seismic structure data reported by Hino et al. [2000] and Takahashi et al. [2004] just to the north suggest that the fore-arc Moho is at an unusually shallow depth of about 20 km. If the latter is correct, the fore-arc mantle corner does not limit downdip rupture. Seno [2005] also found the maximum depth of other thrust seismicity inferred to be on the subduction thrust to extend to over 50 km depth on this margin, much deeper than the estimated fore-arc Moho. However, he found the maximum thrust earthquake depth to be only ~30 km to the north of the 2011 rupture area and to the south off SW Japan, in general agreement with the fore-arc Moho depth. As for Sumatra, rupture off NE Japan appears to extend deeper than the fore-arc mantle corner, but in both cases, this conclusion depends on the fore-arc Moho being exceptionally shallow at 20–25 km. If the fore-arc Moho corner is actually at the more common 35–50 km, there would be agreement with this rupture limit. [45] Lay et al. [2012] in a summary of global great earthquake behavior found that large earthquake displacements most commonly occur to a depth of 35 km. Ruptures of smaller isolated patches on the thrust extend deeper to ~55 km, generating only small magnitude earthquakes. Their results suggest that some smaller thrust events may extend to greater depths than the main displacements in megathrust events. Caution therefore should be used in employing the maximum depth of smaller thrust events for a limit to large displacement rupture and to megathrust earthquake hazard. [46] Another complicating factor that may result in thrust earthquakes deeper than the fore-arc mantle corner is the possibility that some thrust earthquake ruptures occur within subducted sediments, subducted seamounts, or other subduction channel material, isolated from overlying fore-arc mantle serpentinite and talc. More detailed structural studies are needed in the region of the Moho intersection where the downdip rupture limit for both low- and high-frequency energy has been well determined for recent great earthquakes. [47] The thickness of the Cascadia fore-arc crust is quite well determined in several areas by a number of seismic studies, including receiver function analyses, seismic tomography, and wide-angle seismic structure experiments. However, the dip angle of the thrust profile is not so well determined and there remains some debate, especially for the Vancouver Island margin, as to what represents the thrust in seismic structure data [e.g., Hansen et al., 2012]. McCrory et al. [2006, 2012a] provided a review of the thrust constraints. An initial summary of constraints to the location of the fore-arc mantle corner has been provided by McCrory et al. [2012b]. The accuracy is sufficient to show that the fore-arc mantle corner is generally well landward of the seismogenic limit from most of the other constraints (e.g., Figure 14a). The differences in various thrust profile estimates are not significant to a depth of about 30 km [e.g., Wech et al., 2009, their Figure 4] and so are not very important for the seismogenic zone as defined by the geodetic, thermal, and coastal marsh constraints. However, should the seismogenic zone extend deeper, at greater depths, there are significant differences in various estimates of the thrust dip profile. A constraint to the location of the fore-arc mantle corner may come from the updip limit of ETS tremor (discussed later). Wech et al. [2009] suggested that the well-resolved, sharp updip edge of ETS tremor epicenters reflects a change in plate interface coupling properties. This boundary in coupling may be associated with the fore-arc mantle corner (e.g., Figures 13 and 14a), with ETS occurrence being controlled by the position of the fore-arc mantle corner [e.g., Holtkamp and Brudzinski, 2010]. 7. ETS Slow Slip Updip Limit [48] Episodic tremor and slip (ETS) is a remarkable recent discovery, with the most detailed studies in Cascadia and SW Japan [e.g., Rogers and Dragert, 2003; Obara et al., 2004; Schwartz and Rokosky, 2007; Peacock, 2009; Gomberg et al., 2010; Beroza and Ide, 2011]. A number of authors have suggested that ETS may provide a great earthquake downdip rupture limit. The slow slip associated with ETS tremor likely occurs mainly on or just above the subduction thrust, and for Cascadia accommodates much if not all of the plate convergence in the ETS zone [e.g., Dragert and Wang, 2011]. Therefore, there is little or no elastic strain buildup in the ETS zone to be released in great events. If this last conclusion is correct, the slow slip zone represents a downdip limit for significant rupture displacement. The largest uncertainty as to whether most if not all of the plate convergence is so accommodated lies in the amount of slip in small events between the main ETS events [e.g., Wang et al., 2008]. The physical processes involved in ETS remain uncertain, but the slow slip behavior itself indicates that the fault zone has ductile characteristics at that depth. [49] It cannot be excluded that there is some small elastic strain buildup in the ETS zone and that great earthquake rupture extends into the ETS zone with very small displacement. It is recognized that even a small displacement slip in this region could substantially influence the resulting seismic shaking at major cities in Cascadia. If slow slip represents 5540 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Figure 14. (a) Slow slip from inversion of GPS data and tremor locations [Dragert and Wang, 2011] for a portion of the Cascadia margin. The reference locked and transition zones are shown for comparison. The approximate location of the fore-arc mantle corner is from McCrory et al. [2012b]. There is a gap of 50–100 km between the updip limit of slow slip and tremor and most estimates of the downdip limit of significant great earthquake rupture. (b) Distribution of Cascadia tremor 2009–2013 (data from Wech et al. [2009] and Wech [2010], http://www.pnsn.org/tremor), compared to Flück et al. [1997] reference model locked and transition zones. (c) Illustration that the sum of the coseismic downdip transition and the updip transition for the slow slip does not equal the total plate convergence rate over the great earthquake cycle. There must be an intervening region that does not move coseismically or in slow slip events. conditionally stable frictional behavior, is it possible that the slow slip event zone could be pushed to seismic rupture by a large earthquake on the seismogenic portion of the interface. However, the intervening zone between ETS and the main locked/rupture zone estimated from most constraints also would need to rupture coseismically. [50] The horizontal positions of ETS tremor are quite well determined [e.g., Kao et al., 2009; Wech et al., 2009; Gomberg et al., 2010; Dragert and Wang, 2011; Ghosh et al., 2012] (e.g., Figure 14b) although the depths are somewhat uncertain. The location of the slow slip has been estimated through dislocation modeling of geodetic data and its location is also quite well determined but not quite as well as the tremor [e.g., Dragert and Wang, 2011; Wech et al., 2009; McCaffrey, 2009; Schmidt and Gao, 2010; Bartlow et al., 2011]. There is some indication that the slip is slightly seaward of the tremor but this is within the uncertainties (example in Figure 14a). Hawthorne and Rubin [2013] found a short timescale correlation between slow slip and tremor in Cascadia that indicates that they are closely related. [51] Audet et al. [2010] and Ghosh et al. [2012] conclude that the top of the oceanic plate is at a depth of 30–40 km beneath the tremor based on various estimates of plate depth. In their review, Gomberg and the Cascadia 2007 and Beyond Working Group [2010] also concluded that the depth range for Cascadia slow slip is generally 30–40 km. The forearc Moho is approximately at 30 km so the tremor appears to be across and landward of the fore-arc mantle corner but the depth uncertainties for the fore-arc Moho are substantial. In a more conclusive analysis of slow slip from continuous GPS data, Holtkamp and Brudzinski [2010] found that the ETS zone is likely located where the subduction thrust underlies the fore-arc mantle, such that the fore-arc mantle corner is the updip limit of most of the ETS. Peacock [2009] also concluded that the tremor for SW Japan and Cascadia is just below or at the fore-arc mantle corner. A physical composition or structure constraint is supported by the conclusion of Peacock [2009] that tremor occurs at different depths for different subduction zones. He estimated the temperature on the thrust at the depth of the Cascadia tremor to be about 5541 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Figure 15. Gravity anomalies along the Cascadia margin that may delineate the great earthquake rupture extent downdip and possible along-strike segmentation defined by sedimentary basins A–E [after Wells et al., 2003], compared to Flück et al. [1997] reference locked. 510°C, well above the proposed maximum of about 350°C for seismic behavior. The slow slip is therefore 30–50 km landward of the great earthquake rupture zone from the thermal constraint and most of the other constraints for the downdip limit of rupture. [52] A possible explanation for the ETS tremor being associated with the fore-arc mantle corner is that the fluids responsible are channeled upward and seaward beneath the fore-arc mantle to rise above the corner. Very low Poisson’s ratio in the crust above this corner has been observed, suggested by Ramachandran and Hyndman [2011] to result from silica precipitation from the rising fluids. Katayama et al. [2012] also proposed an impermeable barrier at the fore-arc Moho that channels fluids seaward to rise into the crust above the corner producing tremor. [53] If the conclusion that the slow slip of ETS occurs well landward of the great earthquake rupture zone is correct (Figure 14), it means that there is an intervening section of the thrust where long-term plate convergence is accommodated neither in great earthquake rupture nor in slow slip events, but rather in some form of long-term aseismic creep [e.g., Holtkamp and Brudzinski, 2010]. It should be noted that Cascadia has an unusually shallow downdip great earthquake rupture limit, probably because there is a thermal rupture limit for this very hot subduction zone. The more common cool subduction zones often have a much deeper limit that may be at the fore-arc mantle corner and may not have a gap between the great earthquake rupture and ETS slow slip. 8. Geological Associations of Rupture With Offshore Basins [54] There is a common correlation between offshore forearc basins, defined mainly by gravity, and the regions of seismic rupture in past great events globally [Wells et al., 2003; Song and Simons, 2003]. In a related observation, Ruff and Tichelaar [1996] found that the landward limit of great earthquake rupture is often beneath the coast. The coast often defines the landward limit of the basins. This correlation needs to be examined for recent great events such as the 2004 great earthquake of Sumatra, the 2008 great earthquake of Chile, and the 2011 earthquake of NE Japan. [55] Fuller et al. [2006] and Llenos and McGuire [2007] discuss mechanisms to explain this correlation. The largest seismic moment release for the majority of great earthquakes studied elsewhere tended to occur beneath local lows in the gravity field along the margin [e.g., Wells et al., 2003]. Also, the limits of rupture approximately correspond with increases or maxima in the trench parallel gravity field. A physical mechanism has not yet been identified; however, the deep-sea terraces and basins may evolve not only by growth of the outer arc high but also by interseismic subsidence that is not recovered by uplift during great earthquakes. There could be a link between subsidence, subduction erosion, and seismogenesis. Kilometers of late Cenozoic subsidence and crustal thinning above some of the source zones elsewhere are indicated by seismic profiling and drilling. They are thought to be caused by basal subduction erosion [e.g., Scholl and von Huene, 2007] although large amounts of erosion are most commonly argued for sediment deficient subduction zones with high convergence rates [e.g., Stern, 2011] and may not apply to Cascadia. [56] The source zone for the Cascadia 1700 A.D. earthquake contains five semicontinuous basin-centered gravity lows (Figure 15) that may indicate asperities at depth [e.g., Wells et al., 2003]. The gravity gradient marking the inferred downdip limit to large coseismic slip lies offshore, except in northwestern Washington, where the low extends landward beneath the coast. Figure 15 shows the good correlation between the locked and transitions zones defined geodetically and thermally and the gravity-defined basins for Cascadia. The basin landward extents are also in general agreement with the estimated landward limit of rupture in 1700 and earlier great events based on coastal coseismic subsidence [Leonard et al., 2004; 2010]. On other margins, seismic slip has been concluded to be less beneath along-strike structural highs [e.g., Wells et al., 2003] so such highs may delineate rupture variability along the margin with some aseismic slip. 9. Downdip Limit of Small Subduction Thrust Earthquakes [57] The downdip limit of small to intermediate size thrust earthquakes on the subduction thrust provides an important seismogenic zone constraint for many other subduction zones [e.g., Tichelaar and Ruff, 1993]. These thrust events tend to concentrate near the downdip limit of rupture for great earthquakes elsewhere [e.g., Tichelaar and Ruff, 1993; Ruff and Tichelaar, 1996]. [58] However, on the Cascadia margin, only a very few small such events have been detected and in a limited area [Tréhu et al., 2008; Williams et al., 2011] (Figure 16), but 5542 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT displacements in megathrust events and therefore that the maximum depth of smaller thrust events may not define the limit to large megathrust rupture displacement. 10. Landward Limit of Earthquakes on the Nootka Transform Fault Figure 16. Small thrust earthquakes that may define the downdip limit of seismic behavior on the subduction thrust off Oregon [after Williams et al., 2011]. They occurred where thermal model temperatures on the subduction thrust interface are 400–450°C, so they may represent rupture of mafic seamounts that have a higher seismogenic temperature limit than most of the subduction thrust. The downdip limits to the locked (black solid line) and transition (dashed line) zones are from the Flück et al. [1997] reference model. they are at the estimated depth of the subduction thrust and some have been shown to have rupture mechanisms at the dip angle of the thrust. They may be related to local stress concentrations associated with features such as subducted seamounts [Tréhu et al., 2012]. Some are accurately located with ocean bottom seismographs and they may provide an estimate of the downdip seismogenic limit. The estimated source depths are 9–11 km for a M4.9 event and 15–17 km for a M4.8 event. Other small earthquakes are in the range from 10 km to 18 km. These events are within the location uncertainties of the subduction thrust and probably represent seismogenic motion on the Cascadia megathrust. The deeper events are within the transition zone for our reference model [Flück et al., 1997] but well updip of where episodic tremor and slip (ETS) has been documented and well updip of the forearc mantle corner. Williams et al. [2011] concluded that the events occurred at the downdip edge of the locked/transition zone which may be some 30 km farther landward than the Flück et al. model (Figure 16) and suggested that the transition zone extends approximately to the coast. However, an alternative explanation is that these events may have occurred on sheared-off subducted seamounts [Tréhu et al., 2012] or local small blocks of crustal material in the subduction channel with a mafic composition. Such material could have a higher limiting temperature than the normal more felsic sediments and fore-arc crustal rocks on the main thrust contact [e.g., Chen and Molnar, 1983; He et al. 2007] and explain why these thrust earthquakes occur where the temperature is higher than the concluded common seismogenic limit of about 350°C. [59] As noted above, Lay et al. [2012] in a summary of global great earthquake behavior found that large earthquake displacements most commonly occur to a depth of 35 km but that the rupture of smaller isolated patches often extends deeper to ~55 km. Their results suggest that some smaller thrust events may extend to greater depths than the main [60] The left-lateral Nootka transform fault zone lies off central Vancouver Island approximately orthogonal to the margin. The landward downdip extent of the strike-slip seismicity on this fault zone (Figure 17) may provide a limit to the downdip extent of seismic behavior on the overlying subduction thrust. The strike-slip motion is between the main Juan de Fuca plate to the south and Explorer plate complex to the north [Hyndman et al., 1979]. The fault zone is concluded to be the result of quite recent ~3 Ma breakup of the Juan de Fuca plate [e.g., Riddihough, 1984; Braunmiller and Nábělek, 2002]. There is a continuous change in the orientation of this fault zone, resulting in a broad irregular shear zone. The epicenter uncertainties for the Nootka fault zone earthquakes are significant because most seismograph stations are landward. However, a few events have been located with ocean bottom seismographs [Hyndman and Rogers, 1981] and the location bias near the coast is small. Strong seismicity on the Nootka fault zone extends approximately to the coast well landward of the deformation front. Like the small thrust earthquakes off Oregon, the Nootka fault earthquakes extend approximately to the landward limit of the transition zone of the Flück et al. [1997] reference model (Figure 17), where the estimated temperature is about 450°C. [61] It is likely that the Nootka seismicity is thermally limited landward, in a similar manner to the postulated thermal limit for continental strike-slip faults and for the subduction thrust. The seismicity on the subducted Nootka transform fault zone is slightly deeper than the megathrust and therefore may be slightly hotter. The result is a landward seismogenic Figure 17. Seismicity on the Nootka transform fault extending off Vancouver Island that may provide a landward limit of seismic behavior (from Geological Survey of Canada data file), compared to the Flück et al. [1997] reference locked and transition zones. These earthquakes extend to where thermal model temperatures are about 450°C. This higher than normal seismogenic temperature limit may be explained by rupture of mafic ocean crust material that has a higher temperature seismogenic limit than most of the subduction thrust. 5543 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT limit that is slightly seaward. However, the transform seismicity is argued to be only in the upper few kilometers of the oceanic crust [Hyndman and Rogers, 1981; Willoughby and Hyndman, 2005] so not much deeper and not much hotter than the subduction thrust. However, this fault cuts mafic oceanic crustal rocks that likely have a higher maximum seismogenic temperature than for the more felsic composition of the rocks and sediments in contact with the subduction thrust, so there may be a higher maximum seismogenic temperature and therefore the landward limit may be farther landward. For example, Chen and Molnar [1986] estimated the temperature at the base of the seismogenic zone for the fracture zones cutting oceanic lithosphere to be 600–800°C and He et al. [2007] from laboratory data found that seismic behavior could extend to over 510°C for gabbroic oceanic lower crust rocks. This limiting temperature is in general agreement with the landward limit of Nootka seismicity zone (Figure 17) at about 450°C and similar to that inferred for the small thrust earthquakes off Oregon discussed earlier. This constraint provides evidence that significant thrust rupture at least does not extend far beneath the Vancouver Island coast. From this constraint, the maximum megathrust rupture extent may be seaward but is unlikely to be landward. [62] To the south of the Nootka fault zone, there is incoming plate seismicity that extends several tens of kilometers landward of the coast (Figure 17). This seismicity does not continue to the south off Washington and Oregon. It could result from extensional faulting generated by plate bending and steepening landward or due to dehydration fluid pressures [e.g., Abers et al., 2009]. Like the Nooka seismicity and the small events off Oregon, these earthquakes extend to where the estimate temperature is about 450°C which is approximately that expected for mafic ocean crustal rocks. Like the Nootka fault constraint, the maximum megathrust rupture extent may be seaward but is unlikely to be landward of this subducting plate seismicity. 11. Discussion [63] The following is a summary of constraint results: [64] 1. The geodetic constraints for the current locked and transition zones have excellent data and the highest resolution, about ±10 km differences among the various models. The greatest uncertainty is from the degree to which the locked/ transition zones represent the full rupture and the tapering to zero rupture displacement. Recent great earthquakes elsewhere indicate that this is a good approximation. The seismic rupture limit may be slightly seaward of the locked/transition zones because of the postseismic transients that are not included in most locked zone models. [65] 2. Coastal coseismic subsidence provides the least ambiguous constraint to the rupture zone but has quite low resolution of about ±30 km downdip [see Leonard et al., 2010], and there is some ambiguity trade-off of location of rupture and amount of rupture displacement, especially off Oregon where the subsidence is small. There also is the question of how much of the displacement occurs in postseismic deformation. The coastal subsidence constraint agrees well with the geodetic constraint. [66] 3. The thermal constraint has significant uncertainties in both the thermal models and in the critical temperatures (see uncertainties shown in Figure 2), but the results generally agree well with the geodetic constraint. If hydrothermal cooling in the incoming crust is significant, the critical temperatures may be a few tens of kilometers farther landward. [67] 4. The transition in thrust reflection character from thin seismic to thick aseismic has quite high resolution [see Nedimovic et al., 2003] but has a large uncertainty because of the lack of a good calibration where there has been a recent megathrust earthquake. From laboratory and field data, ductile behavior in crustal rocks is expected at roughly 450°C. The reflection transition agrees well with the transition zone thermal limit and the geodetic constraint transition zone limit. [68] 5. The fore-arc mantle corner or Moho limit has considerable uncertainty, about 20 km [e.g., McCrory et al., 2012], but is about 50 km landward of the main rupture indicated by the geodetic models. It roughly agrees with the ETS limit. This seismic rupture limit has been concluded for a number of cold subduction zones elsewhere but there have been several interpretations of thrust earthquakes extending deeper. This constraint is only a maximum depth, and for Cascadia the rupture limit appears to be thermally controlled at a shallower depth than the fore-arc Moho intersection. [69] 6. The ETS slow slip is quite well located to about ±10 km (references given above). The ETS updip limit is where the thrust is at a depth of about 30 km, which corresponds quite well with the fore-arc mantle corner, although the latter is not yet well constrained in some areas of the margin. The ETS updip limit is 30–50 km landward of the geodetic locked/transition zone limit. This constraint is only a maximum depth and allows a shallower rupture limit. For Cascadia, the rupture limit appears to be thermally controlled at a shallower depth. [70] 7. The center of gravity lows beneath the slope and shelf gives a moderate resolution constraint of about ±30 km. There is agreement with the maximum rupture for some but not all great earthquakes elsewhere, but the reason is not yet clear. There is general agreement with the geodetic, coastal paleoseismic, and thermal constraints. [71] 8. The small thrust earthquakes off Oregon that are interpreted to be on the subduction thrust are located with an accuracy of a few kilometers, near the downdip limit of the transition zone of the geodetic and thermal models, deeper than expected for initiation of rupture in felsic rocks. Such events are very rare and may be within seamounts as suggested by Tréhu et al. [2012] or other local mafic material in this area, with a higher temperature limit of about 450°C. [72] 9. The earthquakes on the Nootka transform fault zone are located with an accuracy of a few kilometers. They extend to about the coast of Vancouver Island, i.e., near the landward limit of the transition zone of the geodetic models. The temperature at this location is too hot for seismic initiation in felsic composition rocks. However, the fault cuts mafic oceanic crust that likely has a higher temperature limit than the normal thrust seismic temperature limit. [73] Three of the downdip limits to great earthquake rupture agree within the uncertainties with reasonable resolution of ±10 to ±20 km, i.e., geodetic, coastal paleoseismic, and thermal, and are close to our reference model (e.g., Figures 2 and 8–10). The other limits allow the reference model limit. For major rupture, i.e., ~10 m or 50% of the maximum displacement, the uncertainty is about ±20 km, with a small possibility that the constraints are systematically incorrectly interpreted. The limit for the landward tapering 1–2 m rupture 5544 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT that can still provide strong shaking is less well estimated but appears to be 25–50 km farther landward. This latter limit extends some distance inland of the coast for northern Washington. [74] The small thrust earthquakes off Oregon and the earthquakes on the Nootka fault are interpreted to be subject to a higher temperature limit, ~450°C, than the subduction thrust, ~350°C, because they are interpreted to rupture more mafic rocks, i.e., seamounts and ocean crust, respectively. The change in thrust seismic reflection character from a thin fault to a broad shear zone also occurs at about 450°C as expected from laboratory and field data. The remaining constraints are consistent with thermally limited rupture mainly offshore for this especially hot subduction zone. [75] There are two constraints that provide a fairly strict maximum rupture depth if they are correctly interpreted, albeit with significant uncertainty in position. These are the aseismic fore-arc mantle corner and the updip limit of ETS slow slip that accommodates most if not all plate convergence. The two constraints agree within the large uncertainties. They are a little deeper than the transition zone downdip thermal limit of ~450°C. Both allow a shallower rupture limit and are likely deeper than the actual maximum depth for significant rupture displacement for this hot subduction zone. [76] A surprising result from the analyses of the different downdip rupture limits is that there appears to be a gap of 50–100 km between the downdip limit of significant rupture displacement in great earthquakes and the region of slow slip of ETS. The analysis of slow slip from continuous GPS data by Holtkamp and Brudzinski [2010] supports this conclusion., i.e., a gap between the great earthquake interseismic locked zone and the deeper inter-ETS event locked zone that slips in ETS slow slip events. It is not yet clear how and when this intermediate depth zone moves. Does it move continuously with slow creep with plate convergence accommodated by episodic rupture displacement above and slow slip events below? As an alternative, are there smaller very frequent slow slip events that are too small to be observed geodetically and do not result in tremor? This intermediate depth zone result suggests that the ETS zone does not provide a downdip limit of great earthquake coseismic rupture. Rupture is limited to a shallower depth, although more work is needed to confirm this conclusion. It is noted that for at least some cool subduction zones, the downdip great earthquake rupture limit may coincide with updip limit of ETS slow slip, since both limits may be at the fore-arc mantle corner. [77] The mechanisms responsible for the downdip limit of seismogenesis on subduction thrust faults globally are still not well understood. The thermal limit of subduction thrust rupture appears to apply to hot subduction zones like Cascadia, SW Japan, and Mexico. However, there are only a few subduction zones that are sufficiently hot for the critical temperatures to be reached at a shallow depth to test this hypothesis. The downdip limit of serpentinite/talc at the fore-arc mantle corner agrees with data from a number of subduction zones but is questioned by a number of great earthquakes, including the Sumatra and NE Japan Tohoku M9 earthquakes. Both these earthquakes have been concluded to have some rupture deeper than the fore-arc mantle corner. However, in both cases, this corner is inferred to be exceptionally shallow, 20–25 km, and it is recognized that the depth of the fore-arc Moho tends to be especially difficult to determine. The location of the fore-arc mantle corner is an important need for future study. There also is the question of smaller thrust earthquakes that occur at depths greater than the main rupture displacements. They may represent rupture of seamounts or other mafic crustal material with a higher seismogenic temperature than the main thrust. 12. Seismic Ground Motion From Landward Limits of Rupture [78] For the earthquake shaking at near-coastal cities, models for ground motion with landward distance are needed [e.g., Gregor et al., 2002; Olsen et al., 2008; Atkinson and Macias, 2009]. The landward limit of rupture is one of the most important control of shaking inland for these models. The lack of significant historical earthquakes in Cascadia means it is very difficult to test existing ground motion attenuation models. There are several significant problems in converting the available landward limit constraints for Cascadia to ground motion: [79] 1. The ground motion attenuation with distance relations so far presented used poorly defined distances from the earthquake rupture. Usually, they have used the landward rupture extent for great earthquakes elsewhere as defined by aftershocks or by seismic rupture models which have a large uncertainty. [80] 2. The nature of rupture displacement landward taper, i.e., the transition zone, may be very important, since, as discussed earlier, 1–2 m of displacement can produce very damaging ground motions. This landward taper is not incorporated in most ground motion models. Most great earthquake rupture inversions and models have a transition zone with decreasing rupture displacement landward, corresponding roughly with the estimated tapering of displacement in observed recent great events. For Cascadia with approximately 20 m average full rupture, the 50% or 10 m rupture limit can be estimated with reasonable assurance to within about ±20 km, but the position of the taper to ~2 m (~10%) is less accurately defined 25–50 km farther landward. [81] 3. Within the great earthquake rupture zone, there may be a frequency dependence of seismic energy that varies with downdip distance. More damaging high frequencies may be generated near the landward limit of rupture, even though the rupture displacement may be small there. [82] Acknowledgments. The author would like to thank scientists at the Pacific Geoscience Centre, Geological Survey of Canada for many useful discussions, and a number of authors whose work is discussed in this article for clarifications and additional information. References Abers, G. A., L. S. MacKenzie, S. Rondenay, Z. Zhang, A. G. Wech, and K. C. Creager (2009), Imaging the source region of Cascadia tremor and intermediate-depth earthquakes, Geology, 37, 1119–1122. Atkinson, G. M., and M. Macias (2009), Predicted ground motions for great interface earthquakes in the Cascadia subduction zone, Bull. Seismol. Soc. Am., 99, 1552–1578, doi:10.1785/0120080147. Atwater, B. F. (1987), Evidence for great Holocene earthquakes along the outer coast of Washington State, Science, 236, 942–944. Atwater, B. F., and G. B. Griggs (2012), Deep-sea turbidites as guides to Holocene earthquake history at the Cascadia Subduction Zone— Alternative views for a seismic-hazard workshop: U.S. Geological Survey Open-File Report 2012–1043, 58 p., http://pubs.usgs.gov/of/ 2012/1043/. 5545 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Audet, P., M. G. Bostock, N. I. Christensen, and S. M. Peacock (2009), Seismic evidence for overpressured subducted oceanic crust and megathrust fault sealing, Nature, 457, 76–78. Audet, P., M. G. Bostock, D. C. Boyarko, M. R. Brudzinski, and R. M. Allen (2010), Slab morphology in the Cascadia fore arc and its relation to episodic tremor and slip, J. Geophys. Res., 115, B00A16, doi:10.1029/ 2008JB006053. Bachmann, R., O. Oncken, J. Glodny, W. Seifert, V. Georgieva, and M. Sudo (2009), Exposed plate interface in the European Alps reveals fabric styles and gradients related to an ancient seismogenic coupling zone, J. Geophys. Res., 114, B05402, doi:10.1029/2008JB005927. Bartlow, N. M., S. Miyazaki, A. M. Bradley, and P. Segall (2011), Spacetime correlation of slip and tremor during the 2009 Cascadia slow slip event, Geophys. Res. Lett., 38, L18309, doi:10.1029/2011GL048714. Beroza, G. C., and S. Ide (2011), Slow earthquakes and nonvolcanic tremor, Annu. Rev. Earth Planet. Sci., 39, 271–296, doi:10.1146/annurev-earth040809-152531. Bilek, S. (2007), Influence of subducting topography on earthquake rupture, in The Seismogenic Zone of Subduction Thrust Faults, edited by T. Dixon and C. Moore, pp. 123–146, Columbia Univ. Press, New York. Bilek, S. L., and T. Lay (1999), Rigidity variations with depth along interpolate megathrust faults in subduction zones, Nature, 400, 443–446. Blakely, R. J., T. M. Brocher, and R. E. Wells (2005), Subduction-zone magnetic anomalies and implications for hydrated forearc mantle, Geology, 33, 445–448, doi:10.1130/G21447.1. Blanpied, M. L., D. A. Lockner, and J. D. Byerlee (1991), Fault stability inferred from granite sliding experiments at hydrothermal conditions, Geophys. Res. Lett., 18, 609–612, doi:10.1029/91GL00469. Blanpied, M. L., D. A. Lockner, and J. D. Byerlee (1995), Frictional slip of granite at hydrothermal conditions, J. Geophys. Res., 100, 13,045–13,064, doi:10.1029/95JB00862. Bonner, J. L., D. D. Blackwell, and E. T. Herrin (2003), Thermal constraints on earthquake depths in California, Bull. Seismol. Soc. Am., 93, 2333–2354. Bostock, M. G., R. D. Hyndman, S. Rondenay, and S. M. Peacock (2002), An inverted continental Moho and serpentinization of the forearc mantle, Nature, 417, 536–538, doi:10.1038/417536a. Braunmiller, J., and J. Nábělek (2002), Seismotectonics of the Explorer region, J. Geophys. Res., 107(B10), 2208, doi:10.1029/2001JB000220. Brocher, T. M., T. Parsons, A. M. Tréhu, C. M. Snelson, and M. A. Fisher (2003), Seismic evidence for widespread serpentinized forearc upper mantle along the Cascadia margin, Geology, 31, 267–270, doi:10.1130/ 0091-7613(2003)031<0267:SEFWSF>2.0.CO;2. Burgette, R. J., R. J. Weldon II, and D. A. Schmidt (2009), Interseismic uplift rates for western Oregon and along-strike variation in locking on the Cascadia subduction zone, J. Geophys. Res., 114, B01408, doi:10.1029/ 2008JB005679. Bürgmann, R., and G. Dresen (2008), Rheology of the lower crust and upper mantle: Evidence from rock mechanics, geodesy, and field observations, Annu. Rev. Earth Planet. Sci., 36, 531–567, doi:10.1146/ annurev.earth.36.031207.124326. Calvert, A. J., L. A. Preston, and A. M. Farahbod (2011), Sedimentary underplating at the Cascadia mantle-wedge corner revealed by seismic imaging, Nat. Geosci., 48, 545–548. Chen, W., and P. Molnar (1983), Focal depths of intracontinental and interplate earthquakes and its implications for the thermal and mechanical properties of the lithosphere, J. Geophys. Res., 88, 4183–4214. Chlieh, M., J.-P. Avouac, V. Hjorleifsdottir, T.-R. A. Song, C. Ji, K. S. A. Sladen, H. Hebert, L. Prawirodirdjo, Y. Bock, and J. Galetzka (2007), Coseismic slip and afterslip of the great Mw 9.15 SumatraAndaman earthquake of 2004, Bull. Seismol. Soc. Am., 97, 152–173, doi:10.1785/0120050631. Chlieh, M., J. P. Avouac, K. Sieh, D. H. Natawidjaja, and J. Galetzka (2008), Heterogeneous coupling of the Sumatran megathrust constrained by geodetic and paleogeodetic measurements, J. Geophys. Res., 113, B05305, doi:10.1029/2007JB004981. Clague, J. J., and P. T. Bobrowsky (1994), Evidence for a large earthquake and tsunami 100–400 years ago on western Vancouver Island, British Columbia, Quat. Res., 41, 176–184, doi:10.1006/qres.1994.1019. Clowes, R. M., M. T. Brandon, A. G. Green, C. J. Yorath, A. Sutherland Brown, E. R. Kanasewich, and C. Spencer (1987), LITHOPROBE—Southern Vancouver Island: Cenozoic subduction complex imaged by deep seismic reflections, Can. J. Earth Sci., 24, 31–51. Cole, J., B. Hacker, L. Ratschbacher, J. Dolan, G. Seward, E. Frost, and W. Frank (2007), Localized ductile shear below the seismogenic zone: Structural analysis of an exhumed strike-slip fault, Austrian Alps, J. Geophys. Res., 112, B12304, doi:10.1029/2007JB004975. Cozzens, B. D., and G. A. Spinelli (2012), A wider seismogenic zone at Cascadia due to fluid circulation in subducting oceanic crust, Geology, 40, 899–902, doi:10.1130/G33019.1. Currie, C. A., K. Wang, R. D. Hyndman, and J. He (2004), The thermal effects of steady-state slab-driven mantle flow above a subducting plate: The Cascadia subduction zone and backarc, Earth Planet. Sci. Lett., 223, 35–48. Delouis, B., J. M. Nocquet, and M. Vallée (2010), Slip distribution of the February 27, 2010 Mw=8.8 Maule earthquake, central Chile, from static and high-rate GPS, InSAR, and broadband teleseismic data, Geophys. Res. Lett., 37, L17305, doi:10.1029/2010GL043899. Dessa, J.-X., F. Klingelhoefer, D. Graindorge, C. Andr, H. Permana, M.-A. Gutscher, A. Chauhan, and S. C. Singh (2009), Megathrust earthquakes can nucleate in the forearc mantle: Evidence from the 2004 Sumatra event, Geology, 37, 659–662, doi:10.1130/Gs5653A.1. Dragert, H., and R. D. Hyndman (1995), Continuous GPS monitoring of elastic strain in the northern Cascadia subduction zone, Geophys. Res. Lett., 22, 755–758. Dragert, H., and K. Wang (2011), Temporal evolution of an episodic tremor and slip event along the northern Cascadia margin, J. Geophys. Res., 116, B12406, doi:10.1029/2011JB008609. Escartın, J., G. Hirth, and B. Evans (2001), Strength of slightly serpentinized peridotites: Implications for the tectonics of oceanic lithosphere, Geology, 29, 1023–1026, doi:10.1130/0091-7613(2001)029<1023:SOSSPI>2.0.CO;2. Fagereng, A. (2011), Geology of the seismogenic subduction thrust interface, Geol. Soc. London Spec. Publ., 359, 55–76, doi:10.1144/SP359.4. Flück, P., R. D. Hyndman, and K. Wang (1997), Three-dimensional dislocation model for great earthquakes of the Cascadia subduction zone, J. Geophys. Res., 102, 20,539–20,550, doi:10.1029/97JB01642. Frankel, A. D. (2011), Summary of November 2010 meeting to evaluate turbidite data for constraining the recurrence parameters of great Cascadia earthquakes for the update of the national seismic hazard maps, U.S. Geologic Survey Open-File Report 2011–1310, 13 p., http://pubs. usgs.gov/of/2011/1310/. Fujiwara, T., S. Kodaira, T. No, Y. Kaiho, N. Takahashi, and Y. Kaneda (2011), The 2011 Tohoku-Oki earthquake: Displacement reaching the trench axis, Science, 334, doi:10.1126/science.1211554. Fuller, C. W., S. D. Willett, and M. T. Brandon (2006), Formation of forearc basins and their influence on subduction zone earthquakes, Geology, 34, 65–68, doi:10.1130/G21828.1. Ghosh, A., J. E. Vidale, and K. C. Creager (2012), Tremor asperities in the transition zone control evolution of slow earthquakes, J. Geophys. Res., 117, B10301, doi:10.1029/2012JB009249. Goldfinger, C. (2011), Submarine paleoseismology based on turbidite records, Annu. Rev. Mar. Sci., 3, 35–66, doi:10.1146/annurev-marine120709-142852. Goldfinger, C., et al. (2012), Turbidite event history: Methods and implications for Holocene paleoseismicity of the Cascadia subduction zone, in USGS Professional Paper 1661-F, pp. 184, U.S. Geol. Surv., Reston, VA. Gomberg, J., and the Cascadia 2007 and Beyond Working Group (2010), Slow-slip phenomena in Cascadia from 2007 and beyond: A review, Geol. Soc. Am. Bull., 122, 963–978, doi:10.1130/B30287.1. Gregor, N., W. Silva, I. Wong, and R. Youngs (2002), Ground-motion attenuation relationships for Cascadia subduction zone megathrust earthquakes based on a stochastic finite-fault model, Bull. Seismol. Soc. Am., 92, 1923–1932. Guilbault, J. P., J. J. Clague, and M. Lapointe (1996), Foraminifera evidence for the amount of coseismic subsidence during a late Holocene earthquake on Vancouver Island, west coast of Canada, Quat. Sci. Rev., 5, 913–937, doi:10.1016/S0277-3791(96)00058-3. Gutscher, M.-A., and S. M. Peacock (2003), Thermal models of flat subduction and the rupture zone of great subduction earthquakes, J. Geophys. Res., 108(B1), 2009, doi:10.1029/2001JB000787. Haberland, C., A. Rietbrock, D. Lange, K. Bataille, and S. Hofmann (2006), Interaction between forearc and oceanic plate at the south-central Chilean margin as seen in local seismic data, Geophys. Res. Lett., 33, L23302, doi:10.1029/2006GL028189. Hansen, R. T. J., M. G. Bostock, and N. I. Christensen (2012), Nature of the low velocity zone in Cascadia from receiver function waveform inversion, Earth Planet. Sci. Lett., 337–338, 25–38, doi:10.1016/j.epsl.2012.05.031. Hawkes, A. D., B. P. Horton, A. R. Nelson, C. H. Vane, and Y. Sawai (2011), Coastal subsidence in Oregon, USA, during the giant Cascadia earthquake of AD 1700, Quat. Sci. Rev., 30, 364–376, doi:10.1016/j. quascirev.2010.11.017. Hawthorne, J. C., and A. M. Rubin (2013), Short-time scale correlation between slow slip and tremor in Cascadia, J. Geophys. Res. Solid Earth, 118, 1316–1329, doi:10.1002/jgrb.50103. He, C., Z. Wang, and W. Yao (2007), Frictional sliding of gabbro gouge under hydrothermal conditions, Tectonophysics, 445, 353–362, doi:10.1016/ j.tecto.2007.09.008. Henton, J. A., H. Dragert, K. Wang, H. Kao, and A. Lambert (2010), Investigating northern Cascadia ETS processes with absolute gravity and deformation measurements near Port Renfrew, British Columbia, Am. Geophys. Un., Fall Meeting 2010, San Francisco, Abstract #S23A-2101. 5546 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Hetland, E. A., and M. Simons (2010), Post-seismic and interseismic fault creep II: transient creep and interseismic stress shadows on megathrusts, Geophys. J. Int., 181, 99–112, doi:10.1111/j.1365-246X.2009.04482.x. Heuret, A., S. Lallemand, F. Funiciello, C. Piromallo, and C. Faccenna (2011), Physical characteristics of subduction interface type seismogenic zones revisited, Geochem. Geophys. Geosyst., 12, Q01004, doi:10.1029/ 2010GC003230. Hino, R., S. Ito, H. Shiobara, H. Shirnarnura, T. Sato, T. Kanazawa, J. Kasahara, and A. Hasegawa (2000), Aftershock distribution of the 1994 Sanriku-oki earthquake (Mw 7.7) revealed by ocean bottom seismographic observation, J. Geophys. Res., 105, 21,697–21,710. Hippchen, S., and R. D. Hyndman (2008), Thermal and structural models of the Sumatra subduction zone: Implications for the megathrust seismogenic zone, J. Geophys. Res., 113, B12103, doi:10.1029/2008JB005698. Holtkamp, S., and M. R. Brudzinski (2010), Determination of slow slip episodes and strain accumulation along the Cascadia margin, J. Geophys. Res., 115, B00A17, doi:10.1029/2008JB006058. Hu, Y., and K. Wang (2012), Spherical-Earth finite element model of shortterm postseismic deformation following the 2004 Sumatra earthquake, J. Geophys. Res., 117, B05404, doi:10.1029/2012JB009153. Hu, Y., K. Wang, J. He, J. Klotz, and G. Khazaradze (2004), Three-dimensional viscoelastic finite element model for postseismic deformation of the great 1960 Chile earthquake, J. Geophys. Res., 109, B12403, doi:10.1029/ 2004JB003163. Hyndman, R. D. (2007), The seismogenic zone of subduction thrust faults: What we know and don’t know, in The Seismogenic Zone of Subduction Thrust Faults, edited by T. Dixon and C. Moore, pp. 15–49, Columbia Univ. Press, New York. Hyndman, R. D., and S. M. Peacock (2003), Serpentinization of the forearc mantle, Earth Planet. Sci. Lett., 212, 417–432. Hyndman, R. D., and G. C. Rogers (1981), Seismicity surveys with ocean bottom seismographs off western Canada, J. Geophys. Res., 86, 3867–3880, doi:10.1029/JB086iB05p03867. Hyndman, R. D., and G. C. Rogers (2010), Great earthquakes on Canada’s west coast: A review, Can. J. Earth Sci., 47, 801–820. Hyndman, R. D., and K. Wang (1993), Thermal constraints on the zone of major thrust earthquake failure: The Cascadia subduction zone, J. Geophys. Res., 98, 2039–2060. Hyndman, R. D., and K. Wang (1995), Current deformation and thermal constraints on the zone of potential great earthquakes on the Cascadia subduction thrust, J. Geophys. Res., 100, 22,133–22,154. Hyndman, R. D., R. P. Riddihough, and R. Herzer (1979), The Nootka fault zone—A new plate boundary off western Canada, Geophys. J. R. Astron. Soc., 58, 667–683. Hyndman, R. D., K. Wang, T. Yuan, and G. D. Spence (1993), Tectonic sediment thickening, fluid expulsion, and the thermal regime of subduction zone accretionary prisms: The Cascadia margin off Vancouver Island, J. Geophys. Res., 98, 21,865–21,876. Hyndman, R. D., K. Wang, and M. Yamano (1995), Thermal constraints on the seismogenic portion of the southwestern Japan subduction thrust, J. Geophys. Res., 100, 15,373–15,392, doi:10.1029/95JB00153. Hyndman, R. D., M. Yamano, and D. A. Oleskevich (1997), The seismogenic zone of subduction thrust faults, Island Arc, 6, 244–260. Ito, K. (1990), Regional variations of the cutoff depth of seismicity in the crust and their relation to heat flow and large inland earthquakes, J. Phys. Earth, 38, 5223–5250. Kanda, R. V. S., and M. Simons (2010), An elastic plate model for interseismic deformation in subduction zones, J. Geophys. Res., 115, B03405, doi:10.1029/2009JB006611. Kao, H., S.-J. Shan, H. Dragert, and G. Rogers (2009), Northern Cascadia episodic tremor and slip: A decade of tremor observations from 1997 to 2007, J. Geophys. Res., 114, B00A12, doi:10.1029/2008JB006046. Katayama, I., T. Terada, K. Okazaki, and W. Tanikawa (2012), Episodic tremor and slow slip potentially linked to permeability contrasts at the Moho, Nat. Geosci., 5, 731–734, doi:10.1038/ngeo1559. Kirby, S. H. (1985), Rock mechanics observations pertinent to the rheology of the continental lithosphere and the localization of strain along shear zones, Tectonophysics, 119, 1–27, doi:10.1016/0040-1951(85) 90030-7. Koper, K. D., A. R. Hutko, T. Lay, C. J. Ammon, and H. Kanamori (2011), Frequency-dependent rupture process of the 2011 Mw 9.0 Tohoku earthquake: Comparison of short-period P wave backprojection images and broadband seismic rupture models, Earth Planets Space, 58, 1–4. Kurahashi, S., and K. Irikura (2011), Source model for generating strong ground motions during the 2011 off the Pacific coast of Tohoku Earthquake, Earth, Planet., Space, 63, 571–576. Lay, T., H. Kanamori, C. J. Ammon, K. D. Koper, A. R. Hutko, L. Ye, H. Yue, and T. M. Rushing (2012), Depth-varying rupture properties of subduction zone megathrust faults, J. Geophys. Res., 117, B04311, doi:10.1029/ 2011JB009133. Leonard, L. J., R. D. Hyndman, and S. Mazzotti (2004), Coseismic subsidence in the 1700 great Cascadia earthquake: Coastal estimates versus elastic dislocation models, Geol. Soc. Am. Bull., 116, 655–670, doi:10.1130/ B25369.1. Leonard, L. J., C. A. Currie, S. Mazzotti, and R. D. Hyndman (2010), Rupture area and displacement of past Cascadia great earthquakes from coastal coseismic subsidence, Geol. Soc. Am. Bull., 122, 2079–2096, doi:10.1130/B30108.1. Llenos, A. L., and J. J. McGuire (2007), Influence of fore-arc structure on the extent of great subduction zone earthquakes, J. Geophys. Res., 112, B09301, doi:10.1029/2007JB004944. Loveless, J. P., and B. J. Meade (2011), Spatial correlation of interseismic coupling and coseismic rupture extent of the 2011 MW = 9.0 Tohoku-oki earthquake, Geophys. Res. Lett., 38, L17306, doi:10.1029/2011GL048561. Marone, C., and D. Saffer (2007), Fault friction and the upper transition from seismic to aseismic faulting, in The Seismogenic Zone of Subduction Thrust Faults, edited by T. Dixon and C. Moore, pp. 346–369, Columbia Univ. Press, New York. Mazzotti, S., H. Dragert, J. Henton, M. Schmidt, R. Hyndman, T. James, Y. Lu, and M. Craymer (2003), Current tectonics of northern Cascadia from a decade of GPS measurements, J. Geophys. Res., 108(B12), 2554, doi:10.1029/2003JB002653. Mazzotti, S., A. Lambert, N. Courtier, L. Nykolaishen, and H. Dragert (2007), Crustal uplift and sea level rise in northern Cascadia from GPS, absolute gravity, and tide gauge data, Geophys. Res. Lett., 34, L15306, doi:10.1029/2007GL030283. McCaffrey, R. (2009), Time-dependent inversion of three-component continuous GPS for steady and transient sources in northern Cascadia, Geophys. Res. Lett., 36, doi:10.1029/2008GL036784. McCaffrey, R., and D. Schmidt (2012), Models for the eastern edge of rupture zones of great Cascadia earthquakes: Evidence from GPS and uplift data In, Workshop on Update of Pacific Northwest Portion of the U.S. National Seismic Hazard Maps (NSHMs), March 21–22, Univ. of Washington, Seattle, Wash. McCaffrey, R., A. I. Qamar, R. W. King, R. Wells, G. Khazaradze, C. A. Williams, C. Stevens, J. J. Vollick, and P. C. Zwick (2007), Fault locking, block rotation and crustal deformation in the Pacific Northwest, Geophys. J. Int., 169, 1315–1340, doi:10.1111/j.1365246X.2007.03371.x. McCaffrey, R., L. M. Wallace, and J. Beavan (2008), Slow slip and frictional transition at low temperature at the Hikurangi subduction zone, Nat. Geosci., 1, 316–320, doi:10.1038/ngeo178. McCaffrey, R., R. W. King, S. J. Payne, and M. Lancaster (2013), Active tectonics of northwestern US inferred from GPS-derived surface velocities, J. Geophys. Res. Solid Earth, 118, 709–723, doi:10.1029/2012JB009473. McCrory, P. A., J. L. Blaire, D. H. Oppenheimer, and S. R. Walter (2006), Depth to the Juan de Fuca slab beneath the Cascadia subduction margin —A 3-D model for sorting earthquakes, U.S. Geol. Surv. Data Series 91, Version 1.2, 13. McCrory, P. A., J. L. Blair, F. Waldhauser, and D. H. Oppenheimer (2012a), Juan de Fuca slab geometry and its relation to Wadati-Benioff zone seismicity, J. Geophys. Res., 117, B09306, doi:10.1029/2012JB009407. McCrory, P. A., R. D. Hyndman, and J. L. Blair (2012b), Is there a relationship between Cascadia non-volcanic tremor and the forearc mantle wedge, and the downdip limit of seismogenic rupture? Abstract 1497823, Fall Meeting, Am. Geophys. Un., December 3, 2012. McKenna, J. R., and D. D. Blackwell (2002), Heat sources in subduction zones: Implications for slab seismicity and arc volcanism, in The Cascadia Subduction Zone and Related Subduction Systems, Seismic Structure, Intraslab Earthquakes and Processes, and Earthquake Hazards, edited by S. Kirby, K. Wang, and S. Dunlop, pp. 127–131, U.S. Geol. Surv. Open-File Report 02–328, Geol. Surv. Canada Open File 4350. Miller, M. M., D. J. Johnson, C. M. Rubin, H. Dragert, K. Wang, A. Qamar, and C. Goldfinger (2001), GPS-determination of along-strike variation in Cascadia margin kinematics: Implications for relative plate motion, subduction zone coupling, and permanent deformation, Tectonics, 20, 161–176. Mitchell, C. E., P. Vincent, R. J. Weldon, and M. A. Richards (1994), Present-day vertical deformation of the Cascadia margin, Pacific Northwest, United States, J. Geophys. Res., 99, 12,257–12,277. Moore, D. E., and M. J. Rymer (2007), Talc-bearing serpentinite and the creeping section of the San Andreas fault, Nature, 448, 795–797, doi:10.1038/nature06064. Moore, J., and D. Saffer (2001), Updip limit of the seismogenic zone beneath the accretionary prism of southwest Japan: An effect of diagenetic to lowgrade metamorphic processes and increasing effective stress, Geology, 29, 183–186, doi:10.1130/0091-7613(2001)029. Moore, D. E., D. A. Lockner, S. Ma, R. Summers, and J. D. Byerlee (1997), Strengths of serpentinite gouges at elevated temperatures, J. Geophys. Res., 102, 14,787–14,801. 5547 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Moreno, M., M. Rosenau, and O. Oncken (2010), Maule earthquake slip correlates with pre-seismic locking of Andean subduction zone, Nature, 467, doi:10.1038/nature09349. Nedimovic, M. R., R. D. Hyndman, K. Ramachandran, and G. D. Spence (2003), Reflection signature of seismic and aseismic slip on the northern Cascadia subduction interface, Nature, 24, 416–420. Obara, K., H. Hitoshi, F. Yamamizu, and K. Kasahara (2004), Episodic slow slip events accompanied by non-volcanic tremors in southwest Japan subduction zone, Geophys. Res. Lett., 31, L23602, doi:10.1029/ 2004GL020848. Oleskevich, D. A., R. D. Hyndman, and K. Wang (1999), The updip and downdip limits to great subduction earthquakes: Thermal and structural models of Cascadia, south Alaska, SW Japan, and Chile, J. Geophys. Res., 104, 14,965–14,991. Olsen, K. B., W. J. Stephenson, and A. Geisselmeyer (2008), 3D crustal structure and long-period ground motions from a M 9.0 megathrust earthquake in the Pacific Northwest region, J. Seismol., 12, 145–151. Pacheco, J. F., L. R. Sykes, and C. H. Scholz (1993), Nature of seismic coupling along simple plate boundaries of the subduction type, J. Geophys. Res., 98, 4,133–14,159, doi:10.1029/93JB00349. Peacock, S. (2009), Thermal and metamorphic environment of subduction zone episodic tremor and slip, J. Geophys. Res., 114, B00A07, doi:10.1029/ 2008JB005978. Peacock, S., and R. D. Hyndman (1999), Hydrous minerals in the mantle wedge and the maximum depth of subduction earthquakes, Geophys. Res. Lett., 26, 2517–2520. Peterson, C. D., K. M. Cruikshank, M. E. Darienzo, G. C. Wessen, V. L. Butler, and S. L. Sterling (2012), Coseismic subsidence and paleotsunami run-up records from latest Holocene deposits in the Waatch Valley, Neah Bay, northwest Washington USA: Links to great earthquakes in the northern Cascadia Margin, J. Coastal Res., 29, 157–172. Pollitz, F. F., P. McCrory, D. Wilson, J. Svarc, C. Puskas, and R. B. Smith (2010), Viscoelastic-cycle model of interseismic deformation in the northwestern United States, Geophys. J. Int., 181, 665–696, doi:10.1111/ j.1365-246X.2010.04546.x. Preston, L. A., K. C. Creager, R. S. Crosson, T. M. Brocher, and A. M. Tréhu (2003), Intraslab earthquakes: Dehydration of the Cascadia slab, Science, 302, 1197–1200. Ramachandran, K., and R. D. Hyndman (2011), The fate of fluids released from subducting slab in northern Cascadia, Solid Earth Discuss., 3, 943–962. Reynard, B. (2013), Serpentine in active subduction zones, Lithos, 178, 171–185, doi:10.1016/j.lithos.2012.10.012. Riddihough, R. P. (1984), Recent movements of the Juan de Fuca plate system, J. Geophys. Res., 89, 6980–6994. Rogers, G., and H. Dragert (2003), Episodic tremor and slip on the Cascadia subduction zone: The chatter of silent slip, Science, 300, 1942–1943, doi:10.1126/science.1084783. Ruegg, J. C., A. Rudloff, C. Vigny, R. Madariaga, J. B. de Chabalier, J. Campos, E. Kausel, S. Barrientos, and D. Dimitrov (2009), Interseismic strain accumulation measured by GPS in the seismic gap between Constitución and Concepción in Chile, Phys. Earth Planet. Int., 175, 78–85, doi:10.1016/j.pepi.2008.02.015. Ruff, L. J., and B. W. Tichelaar (1996), What controls the seismogenic plate interface in subduction zones?, in Subduction Top to Bottom, Geophys. Monogr. Ser., vol. 96, edited by G. E. Bebout, W. Scholl, H. Kirby, and P. Platt, pp. 105–112, AGU, Washington, D.C. Sagiya, T., and W. Thatcher (1999), Coseismic slip resolution along a plate boundary megathrust: The Nankai Trough, southwest Japan, J. Geophys. Res., 104, 1111–1129. Savage, J. C. (1983), A dislocation model of strain accumulation and release at a subduction zone, J. Geophys. Res., 88, 4984–4996, doi:10.1029/ JB088iB06p04984. Schmidt, D. A., and H. Gao (2010), Source parameters and time-dependent slip distributions of slow slip events on the Cascadia subduction zone from 1998 to 2008, J. Geophys. Res., 115, B00A18, doi:10.1029/ 2008JB006045. Scholl, D. W., and R. von Huene (2007), Crustal recycling at modern subduction zones applied to the past—Issues of growth and preservation of continental basement crust, mantle geochemistry, and supercontinent reconstruction, in 4-D Framework of Continental Crust, edited by R. D. Hatcher, Jr., M. P. Carlson, J. H. McBride, and J. R. Martínez Catalán, vol 200, pp. 9–32, Geol. Soc. Am. Memoir, Boulder CO doi:10.1130/ 2007.1200(02). Scholz, C. H. (1992), Paradigms or small change in earthquake mechanics, Int. Geophys., 51, 505–517. Scholz, C. H. (1998), Earthquakes and friction laws, Nature, 391, 37–42. Scholz, C. H. (2002), Mechanics of Earthquakes and Faulting, Cambridge Univ. Press, New York. Schwartz, S. Y., and J. M. Rokosky (2007), Slow slip events and seismic tremor at circum-Pacific subduction zones, Rev. Geophys., 45, B00A07, doi:10.1029/2006RG000208. Seno, T. (2005), Variation of downdip limit of the seismogenic zone near the Japanese islands: Implications for the serpentinization mechanism of the forearc mantle wedge, Earth Planet. Sci. Lett., 231, 249–262. Shearer, P., and R. Burgmann (2010), Lessons learned from the 2004 Sumatra-Andaman megathrust rupture, Annu. Rev. Earth Planet. Sci., 38, 103–131. Shillington, D. J., et al. (2011), Constraints on the Aleutian subduction zone from the Shumagin gap to Kodiak asperity from new MCS and OBS data of the ALEUT Project, Abstract T33A-2387 presented at 2011 Fall Meeting, AGU, San Francisco, Calif., 4–9 December, 2011. Sibson, R. H. (1977), Fault rocks and fault mechanisms, J. Geol. Soc., 133, 191–213, doi:10.1144/gsjgs.133.3.0191. Sibson, R. H. (1983), Continental fault structure and the shallow earthquake source, J. Geol. Soc., 140, 741–767, doi:10.1144/gsjgs.140.5.0741. Simoes, M., J. P. Avouac, R. Cattin, and P. Henry (2004), The Sumatra subduction zone: A case for a locked fault zone extending into the mantle, J. Geophys. Res., 109, B10402, doi:10.1029/2003JB002958. Simons, M., et al. (2011), The 2011 magnitude 9.0 Tohoku-Oki earthquake: Mosaicking the megathrust from seconds to centuries, Science, 332, 1421–1425, doi:10.1126/science.1206731. Song, T.-R., and M. Simons (2003), Large trench-parallel gravity variations predict seismogenic behaviour in subduction zones, Science, 301, 630–633. Spinelli, G. A., and K. Wang (2008), Effects of fluid circulation in subducting crust on Nankai margin seismogenic zone temperatures, Geology, 36, 887–890, doi:10.1130/G25145A.1. Stern, C. R. (2011), Subduction erosion: Rates, mechanisms, and its role in arc magmatism and the evolution of the continental crust and mantle, Gondwana Res., 20, 284–308, doi:10.1016/j.gr.2011.03.006. Suito, H., and J. T. Freymueller (2009), A viscoelastic and afterslip postseismic deformation model for the 1964 Alaska earthquake, J. Geophys. Res., 114, B11404, doi:10.1029/2008JB005954. Takahashi, N., S. Kodaira, T. Tsuru, J.-O. Park, Y. Kaneda, K. Suyehiro, H. Kinoshita, S. Abe, M. Nishino, and R. Hino (2004), Seismic structure and seismogenesis off Sanriku region, northeastern Japan, Geophys. J. Int., 159, 129–145, doi:10.1111/ j.1365-246X.2004.02350.x. Tichelaar, B. W., and L. J. Ruff (1991), Seismic coupling along the Chilean subduction zone, J. Geophys. Res., 96, 11,997–12,022, doi:10.1029/ 91JB00200. Tichelaar, B. W., and L. J. Ruff (1993), Depth of seismic coupling along subduction zones, J. Geophys. Res., 98, 2017–2037. Tong, X., et al. (2010), The 2010 Maule, Chile earthquake: Downdip rupture limit revealed by space geodesy, Geophys. Res. Lett., 37, L24311, doi:10.1029/2010GL045805. Tréhu, A. M., J. Braunmiller, and J. L. Nabelek (2008), Probable low-angle thrust earthquakes on the Juan de Fuca-North America plate boundary, Geology, 36, 127–130, doi:10.1130/G24145A.1. Tréhu, A. M., R. J. Blakely, and M. C. Williams (2012), Subducted seamounts and recent earthquakes beneath the central Cascadia forearc, Geology, 40, 103–106, doi:10.1130/G32460.1. Tse, S. T., and J. R. Rice (1986), Crustal earthquake instability in relation to the depth variation of frictional slip properties, J. Geophys. Res., 91, 9452–9472. Verdonck, D. (2005), An inverse dislocation model of surface deformation in western Washington, Tectonophysics, 395, 179–191. Wada, I., and K. Wang (2009), Common depth of slab-mantle decoupling: Reconciling diversity and uniformity of subduction zones, Geochem. Geophys. Geosyst., 10, Q10009, doi:10.1029/2009GC002570. Wada, I., K. Wang, J. He, and R. D. Hyndman (2007), Weakening of the subduction interface and its effects on surface heat flow, slab dehydration and mantle wedge serpentinization, J. Geophys. Res., 113, B04402, doi:10.1029/2007JB005190. Wallace, L. M., et al. (2009), Characterizing the seismogenic zone of a major plate boundary subduction thrust: Hikurangi Margin, New Zealand, Geochem. Geophy. Geosyst., 10, doi:10.1029/2009GC002610. Wang, K. (2007), Elastic and viscoelastic models of crustal deformation in subduction earthquake cycles, in The Seismogenic Zone of Subduction Thrust Faults, edited by T. Dixon, pp. 540–575, Columbia Univ. Press, New York. Wang, P.-L. (2012), Rupture models of the great 1700 Cascadia earthquake based on microfossil paleoseismic observations. MSc Thesis, Univ. of Victoria, Victoria, B.C., 94p. Wang, K., and S. L. Bilek (2011), Do subducting seamounts generate or stop large earthquakes?, Geology, 39, 819–822, doi:10.1130/G31856.1. 5548 HYNDMAN: CASCADIA GREAT EARTHQUAKE RUPTURE LIMIT Wang, K., and T. Dixon (2004), “Coupling” semantics and science in earthquake research, EOS, Trans. Am. Geophys. Un., 85, doi:10.1029/ 2004EO180005. Wang, K., and J. He (1999), Mechanics of low-stress forearcs: Nankai and Cascadia, J. Geophys. Res., 104, 15,191–15,205, doi:10.1029/ 1999JB900103. Wang, K., T. Mulder, G. C. Rogers, and R. D. Hyndman (1995), Case for very low coupling stress on the Cascadia subduction fault, J. Geophys. Res., 100, 12,907–12,918. Wang, K., R. Wells, S. Mazzotti, R. D. Hyndman, and T. Sagiya (2003), A revised dislocation model of interseismic deformation of the Cascadia subduction zone, J. Geophys. Res., 108(B1), 2026, doi:10.1029/2001JB001227. Wang, K., H. Dragert, H. Kao, and E. Roeloffs (2008), Characterizing an “uncharacteristic” ETS event in northern Cascadia, Geophys. Res. Lett., 35, L15303, doi:10.1029/2008GL034415. Wang, K., Y. Hu, and J. He (2012), Deformation cycles of subduction earthquakes in a viscoelastic Earth, Nature, 484, 327–332, doi:10.1038/ nature11032. Wech, A. G. (2010), Interactive tremor monitoring, Seis. Res. Lett., 81(4), 664–669, doi:10.1785/gssrl.81.4.664. Wech, A. G., K. C. Creager, and T. I. Melbourne (2009), Seismic and geodetic constraints on Cascadia slow slip, J. Geophys. Res., 114, B10316, doi:10.1029/2008JB006090. Wells, R. E., R. J. Blakely, Y. Sugiyama, D. W. Scholl, and P. A. Dinterman (2003), Basin-centered asperities in great subduction zone earthquakes: A link between slip, subsidence, and subduction erosion? J. Geophys. Res., 108(B10), 2507, doi:10.1029/2002JB002072. Williams, M. C., A. M. Tréhu, and J. Braunmiller (2011), Seismicity at the Cascadia plate boundary beneath the Oregon continental shelf, Bull. Seismol. Soc. Am., 101, 940–950, doi:10.1785/0120100198. Willoughby, E. C., and R. D. Hyndman (2005), Earthquake rate, slip rate, and the effective seismic thickness for oceanic transform faults of the Juan de Fuca plate system, Geophys. J. Int., 160, 855–868. Wong, I. G., and D. S. Chapman (1990), Deep intraplate earthquakes in the western United States and their relationship to lithospheric temperatures, Bull. Seismol. Soc. Am., 80, 589–599. Yoshioka, S., K. Wang, and S. Mazzotti (2005), Interseismic locking of the plate interface in the northern Cascadia subduction zone, inferred from inversion of GPS data, Earth Planet. Sci. Lett., 231, 239–247. 5549